diff --git a/python/llm/src/bigdl/llm/transformers/models/utils.py b/python/llm/src/bigdl/llm/transformers/models/utils.py
index 22e410d6..ec7b4ad2 100644
--- a/python/llm/src/bigdl/llm/transformers/models/utils.py
+++ b/python/llm/src/bigdl/llm/transformers/models/utils.py
@@ -309,12 +309,13 @@ def use_xmx(x: torch.Tensor, qtype: int):
 
 
 def use_fused_layer_norm(x: torch.Tensor, training: bool):
+    device = get_xpu_device_type(x)
     return (
         not training
         and not x.requires_grad
-        and x.device.type == 'xpu'
+        and device in ["arc", "flex", "pvc", "mtl"]  # fused layer norm cannot run on UHD
         and (
-            get_xpu_device_type(x) not in ["arc", "flex"]
+            device == "mtl"  # fused layer norm conflicts with XMX, so disable it when using XMX
             or x.numel() // x.size(-1) == 1
         )
     )