As an important resource on the demand side,air-conditioning load can flexibly participate in the power system and promote new energy consumption. In the process of air-conditioning load scheduling,there are differences in thermal comfort requirements and scheduling potential among different users. How to satisfy the user differentiated thermal comfort and tap their scheduling potential is a difficult problem to consider. In response to the above problem,the thermal comfort model is firstly introduced to comprehensively quantify the user's thermal comfort experience. Secondly,according to the human parameters in the thermal comfort model,the K-means clustering method is used to divide the user groups with different characteristics of thermal tolerance. On this basis,the air-conditioning direct load control strategy based on user differentiated thermal comfort is proposed and validated by simulation. The example analysis shows that,compared with the traditional strategy,the strategy proposed in this paper fully taps the scheduling potential of different user groups on the basis of completing the consumption task,and guarantees the thermal comfort of them.