Abstract: To ensure the economical and efficient implementation of integrated demand response (IDR), the article proposes three improvements to the traditional user model. First, considering the coupling characteristics among multiple energy household loads during energy-use period, an energy-use comfort coupling matrix is established to improve user's comfort cost model. Furthermore, the reference dependency theory is introduced to describe the coupling characteristics among incentives at different time periods, and an incentive reference matrix is established to optimize the correlation model between user's response power and received incentive prices. Finally, the impact of herd effect caused by bounded rationality on the IDR-participation probability of users is analyzed, and a dynamic participation rate is introduced to characterize the inter-user response behavior coupling characteristics. On the basis of proving the uniqueness of the optimal solution for the proposed model, according to the simulation results, after considering the above factors to improve the user model, the response power deviation and the total cost of MESP are effectively reduced, the total benefit and energy use comfort of users are improved, and the win-win situation between users and MESP is realized. Our research shows that fine-grained modeling of users can ensure that MESPs maintain supply and demand balance in multi-energy systems at a lower cost.