Abstract: With the high penetration of various renewable energy forms, energy storage systems (ESS) can effectively preserve the system's frequency stability, while a distributed ESS can operate under a centralized control strategy in the system automatic generation control (AGC). In this work, we aimed to solve the improper frequency regulation demand allocation presented by current control strategies by suggesting a distributed balancing AGC strategy appropriate for PXP ESS clusters (an ESS type that converts power to other energy forms and then back to power). Specifically, the control strategy relied on a frequency regulation cost function that considers the real-time capacity of ESS, while the distributed nature of our algorithm optimally allocated the system frequency regulation demand aiming to minimize the frequency regulation cost. Our experimental results were highly appealing, demonstrating that the proposed control strategy could break the boundary between power-type and energy type ESS and realize the SoC balancing. Additionally, the distributed algorithm was highly computationally efficient, meeting the plug-and-play ESS requirements. In summary, the strategy in this paper realized the efficient use of PXP ESS cluster in AGC, and improved the frequency stability of the control area.