This study evaluates the suitability of a gridded daily precipitation dataset with 0.5°×0.5° resolution for the upstream basin of Nierji Reservoir. Based on the gridded daily precipitation data and observed data of 17 gauge stations in the studied area, this study analyzed the precision of the gridded data on two aspects: one is the area precipitation value, and the other is 10 days' forecast runoff using three statistical indices, i.e., correlation coefficient (R), relative bias (RE) and root mean square error (RMSE). The results showed that:when using the gridded data to calculate area precipitation, the R, RE and RMSE values between area precipitation with gridded data and the observed area precipitation are 0.7, 16.64%, and 4.09 mm, respectively, which meet the inspection standard of the assessment indices, and the area precipitation with gridded data can describe the precipitation for this area overall though there exists overvaluation. When applying the gridded data to forecast 10 days' runoff, the predicted runoff has a high correlation with the observed runoff. The predicted runoff is slightly larger than the observed runoff in training period and testing period while it underestimates the observed runoff in extrapolation period. The error is relatively large, but it is only 3% different from the predicted runoff value by applying the measured area precipitation. In general, the gridded precipitation data can replace the observed precipitation for this study area.