diff --git a/vllm/v1/worker/xpu_model_runner.py b/vllm/v1/worker/xpu_model_runner.py
index 83ff08371e..5bd2879e69 100644
--- a/vllm/v1/worker/xpu_model_runner.py
+++ b/vllm/v1/worker/xpu_model_runner.py
@@ -61,9 +61,9 @@ def _torch_cuda_wrapper():
     torch.cuda.stream = partial(torch.xpu.stream)
     torch.cuda.set_stream = partial(torch.xpu.set_stream)
 
-    # torch.xpu.Event does not accept the ``blocking`` kwarg that
-    # torch.cuda.Event supports, so drop it here.
-    def _xpu_event(*args, blocking=None, **kwargs):
+    # torch.xpu.Event does not accept the ``blocking`` or ``external`` kwargs
+    # that torch.cuda.Event supports, so drop them here.
+    def _xpu_event(*args, blocking=None, external=None, **kwargs):
         return torch.xpu.Event(*args, **kwargs)
 
     torch.cuda.Event = _xpu_event