A statistical downscaling system for forecasting summer precipitation at stations in Shanxi Province has been developed in this study on the basis of real-time products of numerical prediction from the CGCM/NCC (Coupled Ocean-Atmosphere General Climate Model of National Climate Center) and observational data collected at stations in Shanxi Province. Geopotential height at 500 hPa and sea level pressure in the summer of the current year from the CGCM/NCC simulations are selected as the two predictors; the middle- to lower reaches of the Yangtze River and the tropical central-eastern Pacific are taken as two key regions. The statistical downscaling hindcast of the 2007-2014 summer precipitation over Shanxi Province has improved the prediction compared with the original prediction of the CGCM/NCC. Except for the summer of 2008, the spatial anomaly correlation coefficients (ACC) are statistically significant at the 0.01 level, and the temporal correlation coefficients over most areas of Shanxi Province have also increased significantly with a maximum value of 0.6 by the downscaling scheme. The PS score can be greater than 70. Sample tests indicate that the downscaling method based on canonical correlation analysis (CCA) is applicable in the prediction of summer precipitation anomaly across most of Shanxi with high accuracy and stable performance. Predictions by the downscaling method are much better than the original predictions from the CGCM/NCC.