Abstract:According to the basic process of load forecasting, data preprocessing, model selection and model optimization were separately analyzed. First, it summarizes the traditional data processing methods and briefly introduces new data processing methods. Secondly, the existing short-term load forecasting methods are divided into classical methods, traditional methods and intelligent methods. The application principles of existing forecasting methods are comprehensively analyzed, and the advantages and disadvantages of the forecasting methods are analyzed and compared in detail to improve the accuracy of forecasting. Some new methods have emerged as a result of the move, with the aim of improving prediction accuracy and adapting to a variety of operating conditions. Thirdly, the traditional predictive optimization models are summarized and analyzed, and some existing new optimization models are briefly introduced. The results of these new optimization models are higher than the traditional models, and the advantages of the new optimization model are analyzed. Inadequacies. In the end, the article puts forward the prospect of power system load forecasting in the future. In the short-term load forecasting, it should consider the impact of power market, new energy, and electric vehicle related factors.