Abstract: The advanced adiabatic compressed air energy storage system has characteristics such as pressure attenuation and large-scale changes in operating conditions during the energy release process, which increases the complexity of unit power control. In response to the above issues, this article proposes an adaptive control strategy based on air replenishment and distribution. The PI controller parameters are optimized under all operating conditions through genetic algorithm, and a control parameter library based on output power and storage pressure is established. The online dynamic compensation mechanism is combined to achieve online correction of control parameters. Subsequently, this strategy is used to conduct dynamic characteristic analysis and power step comparison testing of the full energy release cycle. The results show that under two typical conditions, the IAE of the adaptive control strategy is reduced by about 36% and 64% respectively compared to conventional PI control. The proposed adaptive control strategy can achieve fast and smooth power tracking, significantly shorten dynamic adjustment time, and improve the power regulation performance of compressed air energy storage systems during variable operating conditions.