With the explosive growth of power big data, the traditional data processing mode relying on a large amount of manpower and material resources is no longer suitable for the development of modern power systems. Cluster analysis is used to analyze the electricity consumption behavior and daily behaviors of residents in Honghe Prefecture. With the help of the power load data provided by the measurement automation system, and based on user group analysis and recognition, artificial intelligence algorithms such as gray prediction, BP neural network, adaptive BP neural network, PSO algorithm, classified random forest algorithm, and adaptive classification random forest are used to predict the load.Through in?depth research on the consumption habits of local residents and the prediction of power load, theoretical experience and technical support are provided for the application of artificial intelligence to power system customer demand response. The main innovation point is to apply the artificial intelligence algorithm such as adaptive random forest algorithm to the research of customer demand response, reduce the load of the grid during peak period, and adopt the load transfer strategy to reduce the grid operation cost and save energy.