Abstract: With the proposal of the national carbon peak and carbon neutrality goals, urban energy and power consumption, as the main source of carbon dioxide emissions, faces great challenges in its transformation. Due to the great subjective differences in calculating the influencing factors, this paper proposes the influencing factors index system of China's urban energy and power transformation based on the panel data of 296 prefecture-level cities and 4 municipalities in China, based on information entropy. To overcome the limitations of traditional linear statistical models in explaining the influencing factors of urban energy and power transformation, this paper proposes a "mixed model based on SHAP value neural network-individual random effect-time fixed effect", analyzes the influencing factors of China's urban energy and power transformation strategy, expounds the characteristics of China's urban energy and power transformation, and analyzes the differences and causes of the influencing factors of urban energy and power transformation in different regions and different development scale levels, and puts forward targeted development strategy suggestions, which provide a theoretical basis for the formulation of energy and power transformation strategy.