diff --git a/python/paddle/v2/fluid/layers/nn.py b/python/paddle/v2/fluid/layers/nn.py
index e8b4cec6ee638b839e2a7c38e032f74b9cd738ef..3453dd945d558a93a854f99209a6ea8055875d84 100644
--- a/python/paddle/v2/fluid/layers/nn.py
+++ b/python/paddle/v2/fluid/layers/nn.py
@@ -1519,21 +1519,21 @@ def batch_norm(input,
     bias = helper.create_parameter(
         attr=helper.bias_attr, shape=param_shape, dtype=dtype, is_bias=True)
 
-    mean = helper.create_global_variable(
-        name=moving_mean_name,
-        dtype=input.dtype,
+    mean = helper.create_parameter(
+        attr=ParamAttr(
+            name=moving_mean_name, initializer=Constant(0.0), trainable=False),
         shape=param_shape,
-        persistable=True,
-        stop_gradient=True)
-    helper.set_variable_initializer(var=mean, initializer=Constant(0.0))
+        dtype=input.dtype)
+    mean.stop_gradient = True
 
-    variance = helper.create_global_variable(
-        name=moving_variance_name,
-        dtype=input.dtype,
+    variance = helper.create_parameter(
+        attr=ParamAttr(
+            name=moving_variance_name,
+            initializer=Constant(1.0),
+            trainable=False),
         shape=param_shape,
-        persistable=True,
-        stop_gradient=True)
-    helper.set_variable_initializer(var=variance, initializer=Constant(1.0))
+        dtype=input.dtype)
+    variance.stop_gradient = True
 
     # create output
     # mean and mean_out share the same memory