This article proposes a new method for predicting and warning of new energy electricity based on a multi-source prediction data and scene classification recognition algorithm. The study uses ECMWF, GFS, and DERF2.0 multi-source prediction data to improve the accuracy of offshore wind power prediction and disaster warning. In order to overcome the limitations of traditional prediction methods, this study adopts a scene classification recognition method. Based on meteorological, power grid, and environmental data, t…