Abstract: By using the wind speed and direction data collected by the wind measuring equipment, the sliding average scheme can avoid the violent fluctuation of the predicted power caused by the complex and changeable wind conditions. The wind speed-power point was corrected and fitted, and the real output point of the wind turbine was collected to iteratively correct the power curve. A power prediction method is proposed based on the blade Angle, which realizes accurate wind power prediction for mountainous wind farm under complex and variable wind conditions. Through simulation, the validity of the proposed method is verified, and it can be used as a reference for practical engineering.