diff --git a/python-package/xgboost/sklearn.py b/python-package/xgboost/sklearn.py
index 66efb01fa..39ad73415 100644
--- a/python-package/xgboost/sklearn.py
+++ b/python-package/xgboost/sklearn.py
@@ -230,6 +230,9 @@ class XGBModel(XGBModelBase):
             params['missing'] = None  # sklearn doesn't handle nan. see #4725
         if not params.get('eval_metric', True):
             del params['eval_metric']  # don't give as None param to Booster
+        if isinstance(params['random_state'], np.random.RandomState):
+            params['random_state'] = params['random_state'].randint(
+                np.iinfo(np.int32).max)
         return params
 
     def get_xgb_params(self):
diff --git a/tests/python/test_with_sklearn.py b/tests/python/test_with_sklearn.py
index 43c1f2767..a098a97ae 100644
--- a/tests/python/test_with_sklearn.py
+++ b/tests/python/test_with_sklearn.py
@@ -450,6 +450,10 @@ def test_sklearn_random_state():
     clf = xgb.XGBClassifier(random_state=401)
     assert clf.get_xgb_params()['random_state'] == 401
 
+    random_state = np.random.RandomState(seed=403)
+    clf = xgb.XGBClassifier(random_state=random_state)
+    assert isinstance(clf.get_xgb_params()['random_state'], int)
+
 
 def test_sklearn_n_jobs():
     clf = xgb.XGBClassifier(n_jobs=1)