Based on the analyses of relationship between summer precipitation and upper tropospheric temperature, North Atlantic Oscillation (NAO) in the preceding spring, a statistic predictive model for summer precipitation over China is proposed and the forecast results for years 2004－2009 are tested. The results show that the large scale precipitation over China mainly changes with NAO in May, and the meridional dipole pattern of precipitation is associated with NAO in May and the anomalous cooling during 500 hPa－200 hPa in the preceding spring over North China and Northeast China. The NAO influences the summer precipitation over China on both interannual and interdecadal time scales, while the upper tropospheric temperature influences the precipitation mainly on interdecadal time scale. The anomalous condition of spring NAO and upper tropospheric temperature determines the pattern and magnitude of summer precipitation over China via large scale circulation and water vapor transport. A predictive model for summer precipitation over China is proposed based on both this statistic and physical relationship, with the first three principle components (PCs) of summer precipitation as the predictands, and with the first four PCs of upper tropospheric temperature over Asia (10°N－50°N, 60°E－130°E) and the monthly NAO index in the preceding spring as the predictors. The periods of these series with different time scales are also considered to remove the noise and extract useful information. The averaged anomaly correlation coefficient for years 2004－2009 is 0.335, indicting a good skill in forecasting the summer precipitation over China.