Abstract: In recent years, with the steady increase of installed wind capacity in the power grid, accurate wind power forecasting is one of the fundamentals to ensure safe and stable operation of the power system. The current data-driven short-term capacity prediction is mainly based on numerical weather prediction, and the accuracy of this prediction leads to the inability to effectively improve the accuracy of short-term capacity prediction. In response to this issue, this paper proposes a short-term wind power prediction method based on NWP wind speed correction, which utilizes the wind speed trend to construct a new input feature, and establishes a correction model through the coupling effect of the trend and the numerical value in order to correct the NWP wind speed, and applies it to the short-term wind power prediction. Applying the method of this paper to two wind farms in Mengxi, China, the prediction accuracy is improved by 2.3% and 1.2%, respectively. The effectiveness of the method of this paper is verified.