To better adapt PI controller parameters in a microgrid control system to the randomness and fluctuations of renewable energy sources, this paper proposes a four-partition multi-strategy fruit fly optimization algorithm based on adaptive step size for real-time optimization of PI parameters. First, the microgrid overall control system model is established by taking the converters in different micro-source control systems of the wind/ photovoltaic/ gas turbine/ storage energy microgrid as the control objects. The PI parameters are adjusted in real-time based on this model. Then, the fruit fly population is divided into four partitions according to the fitness values of different individuals, and different adaptive updating strategies are designed based on the differences in convergence and diversity of fruit flies in the four partitions. Finally, the algorithm proposed is used to optimize the PI parameters in the control process of each microsource, and compared with three other intelligent algorithms to verify the feasibility and superiority of the proposed algorithm. The simulation results show that the algorithm can make the response of the system converter faster and the output more stable.