With continuous technological innovation, Unmanned Aerial Vehicles (UAVs) have become aerial auxiliary base stations serving ground devices. However, UAVs are frequently unable to accomplish complex computing tasks due to battery capacity constraints. In contrast, wireless power transfer (WPT) technologies offer a promising solution for delivering uninterrupted power to UAVs. This paper studies the Optimizing Multi-UAV Trajectory in Wireless Charging Networks (OMTWCN) problem, involving trajectory optimization algorithm development for multiple wireless-powered UAVs cooperation to achieve independent exploration, cooperation, obstacle avoidance, and charging in wireless charging networks. The goal is to minimize the total UAV path length and the maximum task completion time. To solve this, we propose a three-layer decentralized multi-UAV trajectory optimization architecture. Firstly, we propose a Possible Actions in Current Position (PACP) algorithm to determine possible next actions based on the UAV’s current position. Secondly, we design an algorithm and reward function based on traditional Monte Carlo Tree Search (MCTS), enabling each UAV to independently calculate its next action based on its state. Finally, we propose a Multi-UAV Cooperation Routing Algorithm (MCRA) to control all UAVs’ independent exploration, charging, obstacle avoidance, and return. Simulation results show our algorithm achieves efficient UAV cooperation, reducing total flight distance by 27.75% and maximum task completion time by 29.82% on average.