Abstract: The monthly forecast of wind power generation faces such problems as strong weather information uncertainty and less historical data, so the prediction accuracy is low. Based on the strong seasonal characteristics and the short-term smooth change characteristics of wind power, a technique for expanding historical data of monthly wind power generation was proposed. Based on the extended historical data and the characteristics of forward and backward difference of the medium-term weather forecast information, an integrated entropy weight combination forecasting method based on three prediction algorithms of unit matching, data expansion, and time series is proposed, The validity of the proposed data augmentation method and the accuracy of the comprehensive prediction method are verified by theoretical and example analysis. The proposed data expansion method and comprehensive forecasting method can provide a feasible solution for the monthly power forecast of medium and long-term wind power generation.