To solve the problem of sudden threat, the unmanned aerial vehicle (UAV) replanning the local path, this paper presents a segmented optimization rapidlyexploring random tree (RRT) UAV dynamic path planning algorithm. Firstly, the global algorithm is used to generate the global path, and then the starting and ending points of the local path are determined according to the relevant information of the sudden threat. Finally, the local path is generated by the segmented RRT algorithm. The research shows that the algorithm’s runtime and path cost are reduced by about 10%. For the dynamic path planning, the robustness and realtime performance of the algorithm are strong.