Abstract: The accuracy of the precipitation prediction is enhanced in the mid lower reaches of the Yangtze River in summer, by using the climate factors and based on the analyses of the error filed of the operational numerical prediction model (ONPM) from National Climate Center (NCC). It is found that there are some factors (among the 114 monthly climate factors) that are anomalous before a certain summer. We use the cross-examination to determine the similar threshold by cross-checking the size of an average ACC. Choosing the key anomaly factors impacted the region, picking out the analogue years based on the anomaly degree of the factors, reducing the dimensions of the model’s errors by EOF, making the prediction for the model’s errors using the first three principal components, then the method to aim at the ONPM is proposed based on the factor anomaly and reducing dimensions of the errors of the model. The ACC can be improved up to 0.47 through the method from 0.22 as was corrected by the system’s averaged dependent sample hindcast from 2005 to 2009， which shows good prospect of its operational application.