Abstract: The importance of new energy power supply is becoming increasingly prominent,and the scheduler’s demand for mediumterm wind power prediction is deepening. Regarding the problem of increasing wind speed prediction error with the increase of forecast forward time,the medium term wind power prediction method is proposed based on WRF mode and wind speed error correction. Firstly,WRF mode is used to obtain high spatiotemporal resolution numerical weather forecast data. Then,hybrid Gaussian distribution algorithm is used to quantify the timing rule of the prediction error,and Hidden Markov Model is used to correct the wind speed prediction error. Finally,based on the error-corrected wind speed prediction,the XGBoost prediction model is established for power prediction. Example analysis shows that the proposed error correction algorithm can significantly improve the wind speed prediction accuracy,effectively reduce the medium-term wind power prediction error.