Abstract: To improve the effectiveness of Unmanned Surface Vehicle (USV) path planning in complex marine environments, and to address the issue of low operational efficiency in dynamic environments, a new Dynamic Window Approach (DWA) for USV path planning has been developed. Firstly, the target water area range is determined based on the USV's initial and target positions, and the water area is modelled using a grid system. Secondly, the forces exerted on the USV by wind, waves and currents are calculated and combined with the USV's mass to produce an acceleration value, thereby forming a kinematic model of the USV under the influence of the marine environment. Thirdly, the maximum values of linear and angular acceleration of the USV are determined using the Fossen ship dynamics model, thus obtaining the actual reachable speed set of the USV. Finally, an adaptive weight adjustment algorithm is used to improve DWA and plan the USV path, considering the heading, safe distance and speed factors. Simulation experiments were conducted in the waters near the Zhoushan Islands using the USV "Zhi Kun" to verify the reliability of the model. The results show that, compared with the comparative algorithm, the proposed algorithm performs better in terms of running time, simulation step size, change in bow direction amplitude, change in speed amplitude, and length of the planned path. The improved DWA-planned path ensures the USV can reach its destination safely and quickly, providing a reference for improving the autonomous navigation performance and efficient utilisation of USVs.