In multi-microgrid systems, rising renewable energy penetration causes severe operational volatility and economic challenges, as single-type independent energy storage fails to satisfy dual power and energy requirements. A hybrid shared energy storage architecture is proposed, and a bi-level coordinated optimization model is established based on master-slave nash hybrid game theory, integrating peer-to-peer power trading and shared energy storage agent mechanisms. The upper level is targeted at minimizing storage capacity configuration costs, while the lower level is aimed at optimizing multi-microgrid operational costs. A bi-level hybrid iterative method combining Gurobi and an improved differential evolution algorithm is adopted for solution. Simulation results verify that the proposed strategy significantly improves system economy and renewable energy consumption, effectively coordinates multi-peer-to-peer transactions and shared energy storage scheduling, and provides a novel approach for microgrid cluster optimization in multi-market environments.