Abstract: The dynamic pricing strategy is becoming the generally used regulating means by supply side in the current electricity market. Considering the interactive relationship between retailers and users in the electricity market, a non-cooperative Stackelberg model based on game theory is constructed to study the demand response strategy of household users. The model classifies users to achieve comprehensive consideration of users with different preferences. Meanwhile, the impact of load fluctuations on retailers’ benefits and the effect of users’ satisfaction on the comfort are quantified to improve the utility function on both power supply and consumption side. The backward induction is adopted to solve the Nash equilibrium of the model, and in this way the optimal adjustment of the electricity services under time-of-use price can be realized. Finally, the model is applied to the practical example and the sensitivity of the user category is analyzed. The results show the proposed model has better effect on utility optimization between power supply and consumption side.