Abstract: The traditional power demand forecasting model is not able to meet the demand of the intelligent power system.Therefore,based on the power demand forecasting model of improved evaluation index,the relative ratio method is selected to develop the improved evaluation index system of the power demand forecasting model,and the strategic target layer of the improved power demand evaluation index system is established with the overall target of power demand evaluation,including economic development,population and social development,power grid structure and management level, natural environment,policy and legal environment. According to the improved power demand evaluation index,different scales of wavelet decomposition are implemented to obtain the signal energy fluctuations of different scales in different time periods.The gray model is used to fit and stack the wavelet decomposition results of different scales to obtain the final power demand prediction results. The experimental results show that the model can accurately predict the short-term,medium-term and long-term power demand in the experimental area,and the accuracy of the prediction results is higher than 99%,and the model is highly practicable.