Cruise ships are widely used in water quality monitoring but suffer from path planning problems in a surface water environment. Solutions to path planning problems have higher requirements of path planning distance and path planning time and the ability of real-time obstacles avoidance. This paper introduces a hybrid path planning method to solve the path planning problem when using unmanned cruise ships. First, a model of the surface water environment with unknown districts is established by the grid method. Then, the global path is planned by the A* algorithm based on such a model, and in unknown areas of the model, the artificial potential filed (APF) is employed for local path planning. The problem of obtaining unreachable targets and falling into a local minimum introduced by the APF is improved by using the optimized repulsion potential field function and adding the directional random escape strategy. The A* algorithm with self-learning ability is proposed for secondary planning and cases where a local path cannot be generated. Finally, the optimal path combined with the global path and the local path is smoothed. The simulation results show that the proposed algorithm has better performance than other algorithms from the aspects of distance cost and time cost.