Abstract: In view of the problem of low accuracy in predicting GNSS elevation time series using either the single ARIMA or LSTM model, an ARIMA-LSTM hybrid model is proposed for predicting GNSS elevation time series. The fundamental principle of the ARIMA-LSTM model is analyzed, the establishment and data processing of the model are discussed, and its prediction results are verified through experiments. The experimental results show that in the prediction of GNSS elevation time series, the ARIMA-LSTM model exhibits higher prediction accuracy and robustness compared to either the ARIMA or LSTM model individually.