Abstract: Compressed air energy storage（CAES） technology is one of the key technologies to achieve large-scale wind power access, and has broad prospects in development and application. This paper takes the CAES plant as an important mean to accommodate wind power, and makes it participate in power system optimal scheduling together with generators and demand side resources. Based on the multi-scale forecast error characteristics of wind power, electricity demand and price-based demand response（PDR）, and under the "looking before and behind" principle, a day-ahead and intra-day coordinated dispatch model for the CAES integrated system was proposed. In this model, the interaction mechanism among mass flow rate, power, air pressure and temperature in the CAES plant was considered. The CAES spinning reserve capacity constraints and CAES intra-day scheduling constraints, which can reflect the minute time-scale operation characteristics of CAES, were also considered. The optimal generation schedules and spinning reserve capacity schedules of CAES can be obtained simultaneously through this model. The simulation test was applied on the PJM-5 bus system, and the operation data of the Huntorf CAES plant were used. The simulation results verify the validity of the dispatch strategy.