This paper addresses the issue of autonomous motion planning for unmanned aerial vehicles (UAVs) in maze scenarios. An improved algorithm is proposed based on the EGO-planner, which effectively tackles situations where the UAV’s field of view is entirely obstructed by obstacles. A transition target point is generated by shifting six units from the UAV’s position in the parallel and perpendicular direction to the obstacle, respectively. This process continues until the UAV can bypass the obstacles within its field of view, transitioning through these transition target points before returning to the original local target. Simulation results demonstrate that this algorithmic enhancement effectively guides UAVs to autonomously navigate to designated target points in maze scenarios, offering practicality and feasibility.