As an important demand-side management method, distributed load participation in demand response provides an important means for auxiliary services such as peak-clipping and calming fluctuations of renewable power generation. Air conditioning load is usually viewed as a typical thermostatically related load, and is widely distributed in residential areas. Residential air conditioning load is a kind of demand response resource with broad application prospect. During the peak load period in summer, air conditioning load accounts for more than 40% of urban electricity consumption. Moreover, air conditioning load has certain flexibility and controllability. Reasonable evaluation of its peak clipping potential can effectively promote the safety and economic efficiency of the power grid. This paper proposes an analysis method of air conditioning load demand response potential combined with electricity consumption data and physical model. Subsequently, the switching machine and working state of air-conditioning load are identified, model parameters are calculated, and the dynamic adjustment potential of air-conditioning load in the future period is analyzed. In this paper, the data of electricity consumption in a province of China is used in the case study, and the algorithm is verified by examples.