Abstract:The precondition for the power system to provide reliable, high-quality, economical, and environmentally friendly power for power users is to accurately predict the power load.The dispatching department in the power grid must arrange power generation and power supply plans based on the forecast results of short-term load, so as to optimize resource allocation and improve economic efficiency.Therefore, short-term load forecasting is of great significance. In order to improve the accuracy of load forecasting,the research status on short-term load forecasting method is reviewed comprehensively. Firstly the significance, characteristics and factors affecting the precision of short-term load forecasting are briefly introduced, and the historical development of short-term load forecasting method is comprehensively described. Then the present research situation and problems of various forecasting methods based on data preprocessing and combination prediction are summarized respectively. Finally, the existing problems of the current short-term load forecasting and the probable development trend in the future are pointed out. The conclusion is that after preprocessing historical data, combined with the popular machine learning algorithm, the accuracy of short-term load forecasting can be improved.