diff --git a/apps/desktop/AI_CLASSIFICATION_STATS_FIX.md b/apps/desktop/AI_CLASSIFICATION_STATS_FIX.md
new file mode 100644
index 0000000..15c38bb
--- /dev/null
+++ b/apps/desktop/AI_CLASSIFICATION_STATS_FIX.md
@@ -0,0 +1,135 @@
+# AI分类统计数值溢出问题修复
+
+## 问题描述
+
+用户报告AI分析日志显示异常数据：
+- 总记录数：214（正常）
+- 成功分类：18446744073709552（异常，接近u64最大值）
+
+## 问题分析
+
+### 根本原因
+在 `ai_analysis_log_service.rs` 中，计算成功分类数的逻辑有误：
+
+```rust
+// 错误的计算方式
+successful_classifications: (classification_stats.total_classifications - classification_stats.failed_tasks) as u64,
+```
+
+### 问题详解
+1. `total_classifications` 是分类记录总数（来自 `video_classification_records` 表）
+2. `failed_tasks` 是失败的任务数（来自 `video_classification_tasks` 表）
+3. 这两个数据来源不同，不应该直接相减
+4. 当 `failed_tasks` > `total_classifications` 时，减法结果为负数
+5. 负数转换为 `u64` 时会发生溢出，变成接近 `u64::MAX` 的大数
+
+### 数值分析
+- `u64::MAX` = 18446744073709551615
+- 异常值 = 18446744073709552
+- 差值 = 18446744073709551615 - 18446744073709552 = 63
+
+这表明原始计算结果是 -63，转换为 u64 后变成了这个大数。
+
+## 修复方案
+
+### 1. 扩展 ClassificationStats 结构体
+添加分类记录状态的详细统计：
+
+```rust
+pub struct ClassificationStats {
+    // 原有字段...
+    pub successful_classifications: i32,    // 新增：成功分类记录数
+    pub failed_classifications: i32,        // 新增：失败分类记录数
+    pub needs_review_classifications: i32,  // 新增：需要审核记录数
+}
+```
+
+### 2. 改进数据库查询
+在 `video_classification_repository.rs` 中添加分类记录状态统计：
+
+```rust
+// 获取分类记录状态统计
+let mut stmt = conn.prepare(&format!(
+    "SELECT 
+        SUM(CASE WHEN status = ? THEN 1 ELSE 0 END) as classified,
+        SUM(CASE WHEN status = ? THEN 1 ELSE 0 END) as failed,
+        SUM(CASE WHEN status = ? THEN 1 ELSE 0 END) as needs_review
+     FROM video_classification_records{}", record_where_clause
+))?;
+```
+
+### 3. 修正统计计算逻辑
+使用正确的数据源计算各项统计：
+
+```rust
+Ok(AiAnalysisLogStats {
+    total_records: classification_stats.total_classifications.max(0) as u64,
+    successful_classifications: classification_stats.successful_classifications.max(0) as u64,
+    failed_classifications: classification_stats.failed_classifications.max(0) as u64,
+    needs_review: classification_stats.needs_review_classifications.max(0) as u64,
+    // ...
+})
+```
+
+## 修改文件列表
+
+### 后端文件
+1. `apps/desktop/src-tauri/src/data/models/video_classification.rs`
+   - 扩展 `ClassificationStats` 结构体
+
+2. `apps/desktop/src-tauri/src/data/repositories/video_classification_repository.rs`
+   - 添加分类记录状态统计查询
+   - 更新返回的统计数据
+
+3. `apps/desktop/src-tauri/src/business/services/ai_analysis_log_service.rs`
+   - 修正成功分类数的计算逻辑
+   - 添加数值安全检查
+
+### 前端文件
+1. `apps/desktop/src/types/videoClassification.ts`
+   - 更新 `ClassificationStats` 接口
+
+2. `apps/desktop/src/store/videoClassificationStore.ts`
+   - 更新 `ClassificationStats` 接口
+
+## 安全改进
+
+### 1. 数值溢出防护
+使用 `.max(0)` 确保不会有负数转换为无符号整数：
+
+```rust
+classification_stats.successful_classifications.max(0) as u64
+```
+
+### 2. 数据一致性
+确保统计数据来源一致：
+- 任务相关统计来自 `video_classification_tasks` 表
+- 分类记录相关统计来自 `video_classification_records` 表
+
+### 3. 类型安全
+明确区分有符号和无符号整数的使用场景。
+
+## 测试建议
+
+1. **边界测试**：测试没有分类记录的项目
+2. **数据一致性测试**：验证各项统计数据的逻辑关系
+3. **大数据测试**：测试大量分类记录的性能和准确性
+4. **异常数据测试**：测试数据库中存在异常状态的记录
+
+## 预期结果
+
+修复后，AI分析统计应该显示：
+- 总记录数：214
+- 成功分类：实际成功分类的记录数（≤ 214）
+- 失败分类：实际失败分类的记录数
+- 需要审核：需要人工审核的记录数
+
+所有数值应该在合理范围内，且满足逻辑关系：
+`成功分类 + 失败分类 + 需要审核 = 总记录数`
+
+## 后续优化
+
+1. 添加数据一致性检查
+2. 实现24小时活动统计
+3. 添加更详细的错误日志
+4. 考虑添加数据库索引优化查询性能
diff --git a/apps/desktop/src-tauri/src/business/services/ai_analysis_log_service.rs b/apps/desktop/src-tauri/src/business/services/ai_analysis_log_service.rs
index 8e63c96..27bc645 100644
--- a/apps/desktop/src-tauri/src/business/services/ai_analysis_log_service.rs
+++ b/apps/desktop/src-tauri/src/business/services/ai_analysis_log_service.rs
@@ -265,14 +265,14 @@ impl AiAnalysisLogService {
         let recent_24h_activity = 0; // TODO: 实现24小时内的活动统计
 
         Ok(AiAnalysisLogStats {
-            total_records: classification_stats.total_classifications as u64,
-            successful_classifications: (classification_stats.total_classifications - classification_stats.failed_tasks) as u64,
-            failed_classifications: classification_stats.failed_tasks as u64,
-            needs_review: 0, // TODO: 从数据库查询需要审核的数量
-            total_tasks: classification_stats.total_tasks as u64,
-            completed_tasks: classification_stats.completed_tasks as u64,
-            failed_tasks: classification_stats.failed_tasks as u64,
-            processing_tasks: classification_stats.processing_tasks as u64,
+            total_records: classification_stats.total_classifications.max(0) as u64,
+            successful_classifications: classification_stats.successful_classifications.max(0) as u64,
+            failed_classifications: classification_stats.failed_classifications.max(0) as u64,
+            needs_review: classification_stats.needs_review_classifications.max(0) as u64,
+            total_tasks: classification_stats.total_tasks.max(0) as u64,
+            completed_tasks: classification_stats.completed_tasks.max(0) as u64,
+            failed_tasks: classification_stats.failed_tasks.max(0) as u64,
+            processing_tasks: classification_stats.processing_tasks.max(0) as u64,
             average_confidence: classification_stats.average_confidence,
             average_quality_score: classification_stats.average_quality_score,
             recent_24h_activity,
diff --git a/apps/desktop/src-tauri/src/data/models/video_classification.rs b/apps/desktop/src-tauri/src/data/models/video_classification.rs
index 0e2cbdd..9aef991 100644
--- a/apps/desktop/src-tauri/src/data/models/video_classification.rs
+++ b/apps/desktop/src-tauri/src/data/models/video_classification.rs
@@ -175,6 +175,12 @@ pub struct ClassificationStats {
     pub failed_tasks: i32,
     /// 总分类记录数
     pub total_classifications: i32,
+    /// 成功分类记录数
+    pub successful_classifications: i32,
+    /// 失败分类记录数
+    pub failed_classifications: i32,
+    /// 需要审核记录数
+    pub needs_review_classifications: i32,
     /// 平均置信度
     pub average_confidence: f64,
     /// 平均质量评分
diff --git a/apps/desktop/src-tauri/src/data/repositories/video_classification_repository.rs b/apps/desktop/src-tauri/src/data/repositories/video_classification_repository.rs
index ec9b392..a5a1fcb 100644
--- a/apps/desktop/src-tauri/src/data/repositories/video_classification_repository.rs
+++ b/apps/desktop/src-tauri/src/data/repositories/video_classification_repository.rs
@@ -536,6 +536,27 @@ impl VideoClassificationRepository {
             ))
         })?;
 
+        // 获取分类记录状态统计
+        let mut stmt = conn.prepare(&format!(
+            "SELECT
+                SUM(CASE WHEN status = ? THEN 1 ELSE 0 END) as classified,
+                SUM(CASE WHEN status = ? THEN 1 ELSE 0 END) as failed,
+                SUM(CASE WHEN status = ? THEN 1 ELSE 0 END) as needs_review
+             FROM video_classification_records{}", record_where_clause
+        ))?;
+
+        let classified_json = serde_json::to_string(&ClassificationStatus::Classified).unwrap();
+        let failed_json = serde_json::to_string(&ClassificationStatus::Failed).unwrap();
+        let needs_review_json = serde_json::to_string(&ClassificationStatus::NeedsReview).unwrap();
+
+        let record_status_stats = stmt.query_row([&classified_json, &failed_json, &needs_review_json], |row| {
+            Ok((
+                row.get::<_, Option<i32>>(0)?.unwrap_or(0),
+                row.get::<_, Option<i32>>(1)?.unwrap_or(0),
+                row.get::<_, Option<i32>>(2)?.unwrap_or(0),
+            ))
+        })?;
+
         Ok(ClassificationStats {
             total_tasks: task_stats.0,
             pending_tasks: task_stats.1,
@@ -543,6 +564,9 @@ impl VideoClassificationRepository {
             completed_tasks: task_stats.3,
             failed_tasks: task_stats.4,
             total_classifications: record_stats.0,
+            successful_classifications: record_status_stats.0,
+            failed_classifications: record_status_stats.1,
+            needs_review_classifications: record_status_stats.2,
             average_confidence: record_stats.1,
             average_quality_score: record_stats.2,
         })
diff --git a/apps/desktop/src/store/videoClassificationStore.ts b/apps/desktop/src/store/videoClassificationStore.ts
index 534058f..b85bd12 100644
--- a/apps/desktop/src/store/videoClassificationStore.ts
+++ b/apps/desktop/src/store/videoClassificationStore.ts
@@ -60,6 +60,9 @@ export interface ClassificationStats {
   completed_tasks: number;
   failed_tasks: number;
   total_classifications: number;
+  successful_classifications: number;
+  failed_classifications: number;
+  needs_review_classifications: number;
   average_confidence: number;
   average_quality_score: number;
 }
diff --git a/apps/desktop/src/types/videoClassification.ts b/apps/desktop/src/types/videoClassification.ts
index f3cf0a5..e1bb911 100644
--- a/apps/desktop/src/types/videoClassification.ts
+++ b/apps/desktop/src/types/videoClassification.ts
@@ -121,7 +121,10 @@ export interface ClassificationStats {
   completed_tasks: number;
   failed_tasks: number;
   total_classifications: number;
-  average_confidence_score: number;
+  successful_classifications: number;
+  failed_classifications: number;
+  needs_review_classifications: number;
+  average_confidence: number;
   average_quality_score: number;
 }