This paper employs the GARCH-DJI-MIDAS model that is proposed by Pan et al. (2020) to investigate the relationship between economic policy uncertainty (EPU) and stock market volatility. An empirical application to China's stock indices shows that EPU has a significantly negative impact on stock market volatility. Moreover, jump intensity is time-varying and high persistence. In addition, the GARCH-DJI-MIDAS-EPU model outperforms the GARCH, GARCH-MIDAS and GARCH-MIDAS-EPU models in in-sample fitting. Meanwhile, out-of-sample analysis based on four loss functions and the model confidence set (MCS) test suggests that the GARCH-DJI-MIDAS-EPU model yields more accurate out-of-sample volatility forecasts than competing models, suggesting introducing the EPU and jump dynamics can offer high prediction accuracy of stock market volatility.