As one of the important flexible load resources, air conditioning participates in power grid regulation through load aggregators. This is of significance in improving the load characteristics of a power grid in summer. However, aggregators cannot maximize the adjustable potential of air conditioning through group control, and it has a certain impact on user comfort. From the point of view of load aggregators, the temperature setting values are used to control the user air conditioning load in order to maintain the diversity of user regulation behavior. In order to maximize the potential of load adjustment while ensuring user comfort, a two-level optimization model of day ahead scheduling for load aggregators based on user air conditioning load temperature control is proposed. The upper layer of the model takes the maximum benefit of the load aggregator as the optimization objective, and the lower layer considers the difference of user comfort and the minimum level of overall user discomfort as the optimization objective. The particle swarm integer programming algorithm is used to obtain the adjustment of the set temperature of a single piece of air conditioning equipment. The simulation results show that the day ahead scheduling strategy can provide certain peak clipping capability to the power grid while fully tapping the adjustable potential of the air-conditioning load to ensure that the load aggregator obtains the maximum benefit and the user comfort level is higher.This work is supported by the National Natural Science Foundation of China (No. 52007126 and No. U2166209).