diff --git a/python/paddle/fluid/contrib/slim/quantization/quant2_int8_mkldnn_pass.py b/python/paddle/fluid/contrib/slim/quantization/quant2_int8_mkldnn_pass.py
index 6557a0aec010137d4417cfe5329bff41652698d9..f8de55ee3caea9f4f7766bfbab238e03a9ca9926 100644
--- a/python/paddle/fluid/contrib/slim/quantization/quant2_int8_mkldnn_pass.py
+++ b/python/paddle/fluid/contrib/slim/quantization/quant2_int8_mkldnn_pass.py
@@ -448,13 +448,14 @@ class Quant2Int8MkldnnPass(object):
             graph = self._apply_pass(graph, 'fc_act_mkldnn_fuse_pass')
         graph = self._apply_pass(graph,
                                  'matmul_transpose_reshape_mkldnn_fuse_pass')
+        graph = self._apply_pass(graph,
+                                 'matmul_elementwise_add_mkldnn_fuse_pass')
+        graph = self._apply_pass(graph, 'matmul_activation_mkldnn_fuse_pass')
         graph = self._apply_pass(graph, 'batch_norm_act_fuse_pass')
         graph = self._apply_pass(graph, 'softplus_activation_mkldnn_fuse_pass')
         graph = self._apply_pass(graph, 'scale_matmul_fuse_pass')
         graph = self._apply_pass(graph,
                                  'reshape_transpose_matmul_mkldnn_fuse_pass')
-        graph = self._apply_pass(graph,
-                                 'matmul_elementwise_add_mkldnn_fuse_pass')
         # the following pass should be the last one since it will work on all fused ops.
         graph = self._apply_pass(graph, 'runtime_context_cache_pass')
         return graph