Nowadays, dynamical climate models are inefficient in meeting the real needs of climate prediction. An effective method is the combination of dynamical and statistical models. This combination integrates large-scale circulation information from the dynamical models into the statistical model to improve the prediction skill. On the basis of the higher prediction skill for the large-scale summer circulation variable of climate models and the significant relationship between the preceding ENSO signal and summer precipitation in China, a hybrid statistical downscaling prediction method for summer precipitation anomaly prediction in China was proposed in this paper. Cross validation of seasonal prediction for summer precipitation in China was performed, and results showed that the downscaling method improved the multi-year average of anomaly correlation coefficient significantly. In real application, the average PS score reached 71.5/72.7 during 2013–2018/2015–2018, which is higher than that of the original model and the operational predictions issued by the Beijing Climate Center. This statistical downscaling model, which has stable predictive skill, is one of the effective references for operational seasonal prediction in China.