Using year-to-year increments of soil moisture in nine key regions over Eurasia as predictors, statistical prediction models were developed. These models were based on Barnett-Preisendorfer Canonical Correlation Analysis (BP-CCA) and were combined with the Ensemble Canonical Correlation analysis (ECC) method to predict the year-to-year increments of summer precipitation over eastern China, and thus obtain the prediction of summer precipitation. Specifically, data during the epochs of 1980-2004 and 2005-2014 were used to perform historical prediction and independent sample tests, respectively. First, single factor prediction models of the nine predictors were built using the BP-CCA method. Then, ensemble prediction models were developed using the ECC method, based on different combinations of the nine predictors, and the scores for the region with predictive skill in the independent sample test were also calculated. The results showed that the combinations of different predictors had different predictive skill for summer precipitation in China. The soil moisture in the eastern European Plain, in the area north of Lake Baikal, in Hetao region, and in the area south of the Yangtze River, had a good predictive effect for summer precipitation in North China. The soil moisture in the area north of Balkhash Lake, in Northwest China, in Hetao region, and in the area south of the Yangtze River, had a good predictive effect for summer precipitation in the Yangtze-Huaihe region. The soil moisture in the eastern European Plain, in the area north of Balkhash Lake, and in Hetao region, had high predictive skill for summer precipitation in South China. The precipitation change trends predicted by three models were consistent with observations in the corresponding regions. Prediction scores all exceeded 70 points, and anomaly correlation coefficients were all positive. The study shows that soil moisture contains useful signals for summer precipitation in China, and can be considered for application in summer precipitation prediction operations.