Path planning is one of the key technologies to realize the hidden navigation of underwater vehicles during long-haul. Path planning efficiency and accuracy are at the core of submarine track planning. Combining the navigation task with the geomagnetic map adaptability, the optimal path between the starting point and the target point is searched in the target space. The underwater geomagnetic navigation path planning model is established, and the principle and implementation method of Dijkstra algorithm are analyzed. An underwater geomagnetic navigation path planning model is established, and the Dijkstra algorithm is used for underwater geomagnetic navigation path planning. Combining different local windows in the adaptation area, the path planning calculation time and track cost are optimized. The simulation analyzes the influence of different local windows on the path planning in the adaptation area. The experiment results demonstrate that the Dijkstra algorithm can effectively find the optimal path that satisfies the constraints.