The parameter tuning problem of complex control systems is often a significant factor that troubles controller design. While most fixed-wing aircraft use simple PID control methods, tuning multiple loops and groups of parameters still involves a considerable amount of work. In this paper, based on the above phenomenon, a multi-PID control parameter overall tuning method based on reinforcement learning is designed for the DEP fixed-wing aircraft. The research sets up an environment based on the established DEP aircraft simulation model and designs a connector between the simulation model and the reinforcement learning environment. The proposed parameter tuning method can simultaneously tune the control parameters of all control loops of the fixed-wing aircraft and achieve good response effects.