diff --git a/flaml/model.py b/flaml/model.py
index 09cef1b34..30b9d0969 100644
--- a/flaml/model.py
+++ b/flaml/model.py
@@ -254,7 +254,7 @@ class BaseEstimator:
         """
         if self._model is not None:
             X = self._preprocess(X)
-            return self._model.predict(X)
+            return self._model.predict(X, **kwargs)
         else:
             logger.warning(
                 "Estimator is not fit yet. Please run fit() before predict()."
@@ -277,7 +277,7 @@ class BaseEstimator:
         assert self._task in CLASSIFICATION, "predict_proba() only for classification."
 
         X = self._preprocess(X)
-        return self._model.predict_proba(X)
+        return self._model.predict_proba(X, **kwargs)
 
     def score(self, X_val: DataFrame, y_val: Series, **kwargs):
         """Report the evaluation score of a trained estimator.
@@ -312,7 +312,7 @@ class BaseEstimator:
                 )
             else:
                 X_val = self._preprocess(X_val)
-                metric = kwargs.get("metric", None)
+                metric = kwargs.pop("metric", None)
                 if metric:
                     y_pred = self.predict(X_val, **kwargs)
                     if is_min_metric(metric):
@@ -1321,7 +1321,7 @@ class XGBoostEstimator(SKLearnEstimator):
         if not issparse(X):
             X = self._preprocess(X)
         dtest = xgb.DMatrix(X)
-        return super().predict(dtest)
+        return super().predict(dtest, **kwargs)
 
     @classmethod
     def _callbacks(cls, start_time, deadline):
@@ -1823,7 +1823,7 @@ class Prophet(SKLearnEstimator):
             )
         if self._model is not None:
             X = self._preprocess(X)
-            forecast = self._model.predict(X)
+            forecast = self._model.predict(X, **kwargs)
             return forecast["yhat"]
         else:
             logger.warning(
@@ -1835,7 +1835,7 @@ class Prophet(SKLearnEstimator):
         from sklearn.metrics import r2_score
         from .ml import metric_loss_score
 
-        y_pred = self.predict(X_val)
+        y_pred = self.predict(X_val, **kwargs)
         self._metric = kwargs.get("metric", None)
         if self._metric:
             return metric_loss_score(self._metric, y_pred, y_val)
@@ -1916,10 +1916,10 @@ class ARIMA(Prophet):
                     X = self._preprocess(X.drop(columns=TS_TIMESTAMP_COL))
                     regressors = list(X)
                     forecast = self._model.predict(
-                        start=start, end=end, exog=X[regressors]
+                        start=start, end=end, exog=X[regressors], **kwargs
                     )
                 else:
-                    forecast = self._model.predict(start=start, end=end)
+                    forecast = self._model.predict(start=start, end=end, **kwargs)
             else:
                 raise ValueError(
                     "X needs to be either a pandas Dataframe with dates as the first column"
@@ -2120,7 +2120,7 @@ class TS_SKLearn(SKLearnEstimator):
                     ) = self.hcrystaball_model._transform_data_to_tsmodel_input_format(
                         X.iloc[:i, :]
                     )
-                    preds.append(self._model[i - 1].predict(X_pred)[-1])
+                    preds.append(self._model[i - 1].predict(X_pred, **kwargs)[-1])
                 forecast = DataFrame(
                     data=np.asarray(preds).reshape(-1, 1),
                     columns=[self.hcrystaball_model.name],
@@ -2131,7 +2131,7 @@ class TS_SKLearn(SKLearnEstimator):
                     X_pred,
                     _,
                 ) = self.hcrystaball_model._transform_data_to_tsmodel_input_format(X)
-                forecast = self._model.predict(X_pred)
+                forecast = self._model.predict(X_pred, **kwargs)
             return forecast
         else:
             logger.warning(