diff --git a/astrbot/core/utils/dify_api_client.py b/astrbot/core/utils/dify_api_client.py
index fbc8a7e59..30ac08de1 100644
--- a/astrbot/core/utils/dify_api_client.py
+++ b/astrbot/core/utils/dify_api_client.py
@@ -35,19 +35,30 @@ class DifyAPIClient:
                 text = await resp.text()
                 raise Exception(f"chat_messages 请求失败：{resp.status}. {text}")
             
+            buffer = ""
             while True:
-                data = await resp.content.read(8192) # 防止数据过大导致高水位报错
-                if not data:
+                # 保持原有的8192字节限制，防止数据过大导致高水位报错
+                chunk = await resp.content.read(8192)
+                if not chunk:
                     break
-                if not data.strip():
-                    continue
-                elif data.startswith(b"data:"):
-                    try:
-                        json_ = json.loads(data[5:])
-                        yield json_
-                    except BaseException:
-                        pass
                     
+                buffer += chunk.decode('utf-8')
+                blocks = buffer.split('\n\n')
+                
+                # 处理完整的数据块
+                for block in blocks[:-1]:
+                    if block.strip() and block.startswith('data:'):
+                        try:
+                            json_str = block[5:]  # 移除 "data:" 前缀
+                            json_obj = json.loads(json_str)
+                            yield json_obj
+                        except json.JSONDecodeError as e:
+                            logger.error(f"JSON解析错误: {str(e)}")
+                            logger.error(f"原始数据块: {json_str}")
+                
+                # 保留最后一个可能不完整的块
+                buffer = blocks[-1] if blocks else ""
+
     async def workflow_run(
         self,
         inputs: Dict,
@@ -66,20 +77,32 @@ class DifyAPIClient:
         ) as resp:
             if resp.status != 200:
                 text = await resp.text()
-                raise Exception(f"chat_messages 请求失败：{resp.status}. {text}")
+                raise Exception(f"workflow_run 请求失败：{resp.status}. {text}")
+            
+            buffer = ""
             while True:
-                data = await resp.content.read(8192) # 防止数据过大导致高水位报错
-                if not data:
+                # 保持原有的8192字节限制，防止数据过大导致高水位报错
+                chunk = await resp.content.read(8192)
+                if not chunk:
                     break
-                if not data.strip():
-                    continue
-                elif data.startswith(b"data:"):
-                    try:
-                        json_ = json.loads(data[5:])
-                        yield json_
-                    except BaseException:
-                        pass
                     
+                buffer += chunk.decode('utf-8')
+                blocks = buffer.split('\n\n')
+                
+                # 处理完整的数据块
+                for block in blocks[:-1]:
+                    if block.strip() and block.startswith('data:'):
+                        try:
+                            json_str = block[5:]  # 移除 "data:" 前缀
+                            json_obj = json.loads(json_str)
+                            yield json_obj
+                        except json.JSONDecodeError as e:
+                            logger.error(f"JSON解析错误: {str(e)}")
+                            logger.error(f"原始数据块: {json_str}")
+                
+                # 保留最后一个可能不完整的块
+                buffer = blocks[-1] if blocks else ""
+
     async def file_upload(
         self,
         file_path: str,