With the increasing intelligence and autonomy of Unmanned Aerial Vehicles (UAVs), autonomous path planning has become one of their core technologies. When UAVs perform missions in complex and uncertain flight environments, low-confidence environmental data can lead to reduced planning accuracy, potentially leading to collisions with obstacles or entry into hazardous zones. An improved global static path planning algorithm is proposed and an autonomous UAV path planning method that combines a local dynamic path correction strategy is presented. By fully utilizing real-time environmental perception results to locally optimize the static trajectory, the proposed method enhances the UAV’s autonomous path planning capability in complex and dynamic environments. Experimental results confirm that our method overcomes the poor adaptability of conventional global static path planning algorithms while enhancing the computational efficiency and success rate of local planning techniques. A theoretical foundation and technical support for UAV flight safety in complex environments is provided.