Abstract: An ensemble learning model algorithm based on Stacking strategy is proposed. Through the evaluation stage and the integration stage of the basic model algorithms, K basic models are successfully selected, and the integration work of models based on model ensemble strategy is completed to achieve the integrated prediction model based on Stacking strategy. The integrated model is applied to predict the train stopping error of practical cases, and the prediction results are verified. The verification results show that the ensemble learning model algorithm based on Stacking strategy has high training efficiency and excellent prediction accuracy, which is a strong advantage compared to other conventional models.