diff --git a/python/sglang/srt/layers/dp_attention.py b/python/sglang/srt/layers/dp_attention.py
index b7feccdb6..3bae7e0c3 100644
--- a/python/sglang/srt/layers/dp_attention.py
+++ b/python/sglang/srt/layers/dp_attention.py
@@ -51,7 +51,12 @@ class DpPaddingMode(IntEnum):
         return self == DpPaddingMode.SUM_LEN
 
     @classmethod
-    def get_dp_padding_mode(cls, global_num_tokens: List[int]) -> DpPaddingMode:
+    def get_dp_padding_mode(
+        cls, is_extend_in_batch, global_num_tokens: List[int]
+    ) -> DpPaddingMode:
+        if is_extend_in_batch:
+            return DpPaddingMode.SUM_LEN
+
         # we choose the mode that minimizes the communication cost
         max_len = max(global_num_tokens)
         sum_len = sum(global_num_tokens)
diff --git a/python/sglang/srt/model_executor/forward_batch_info.py b/python/sglang/srt/model_executor/forward_batch_info.py
index e343d6b4f..017b5863c 100644
--- a/python/sglang/srt/model_executor/forward_batch_info.py
+++ b/python/sglang/srt/model_executor/forward_batch_info.py
@@ -686,7 +686,9 @@ class ForwardBatch:
                 (global_num_tokens[i] - 1) // attn_tp_size + 1
             ) * attn_tp_size
 
-        dp_padding_mode = DpPaddingMode.get_dp_padding_mode(global_num_tokens)
+        dp_padding_mode = DpPaddingMode.get_dp_padding_mode(
+            self.is_extend_in_batch, global_num_tokens
+        )
         self.dp_padding_mode = dp_padding_mode
 
         if dp_padding_mode.is_max_len():