Abstract: To address the dynamic obstacle avoidance challenge in unmanned surface vehicle (USV) navigation within natural environments, this study proposed an enhanced RRT*-DWA strategy based on a greedy algorithm. Firstly, a comprehensive analysis of prevalent global and local path planning approaches; secondly, the development of USV motion analysis and dynamic obstacle simulation environment modeling; finally, the integration and validation of global RRT* and local DWA methods, complemented by greedy pruning optimization to enhance real-time responsiveness and path smoothing. In dynamic scenario path planning for USV, the improved RRT* algorithm demonstrates significant performance enhancements, achieving an 82.21% reduction in average runtime, a 5.66% decrease in average path length, and an 81.67% reduction in node count, while reducing USV navigation distance by 1.78%. The results show that the enhanced algorithm effectively improves USV's obstacle avoidance and path planning capabilities in dynamic environments.