Abstract:This paper addresses the problem of intelligent control of aircraft overload and attitude in complex and changing environments. It proposes a distributed intelligent agent control method based on the Soft Actor-Critic (SAC) algorithm, establishes a framework of distributed efficient environment interaction for deep reinforcement learning algorithms, and designs an intelligent control algorithm system for aircraft overloadand attitude. This approach increases the scale and distribution of training data for reinforcement learning algorithms, thereby improving the performance and robustness of aircraft control algorithms. Experimental results in simulated environments demonstrate that the trained intelligent agents effectively control overloadsandattitudes in unmanned aerial vehicle simulations. The distributed SAC algorithm outperforms the original SAC algorithm in controlling unmanned aerial vehicles in simulation scenarios.