A method for autonomous path planning for disaster monitoring such as Fukushima nuclear power station accident in 2011 using a small rotorcraft UAV is proposed. In this method, a rotorcraft can search around damaged buildings without specifying the waypoints by an operator. Airspace including obstacles is assumed to be composed of 3D cells, and the vertexes of 3D cells become the candidates for waypoints. The waypoints through which the UAV can pass are obtained as flyable waypoints by searching and avoiding the position of 3D cells which include the objects inside them using laser scanning. The UAV first chooses the waypoints suitable for taking images from the flyable waypoints using evolutionary computation, and then generates 3D flight path using A*-EC hybrid path planning method. A*-EC hybrid path planning method is a technique for generating flight path quickly considering obstacle avoidance. Path planning problem is solved while the number of waypoints and final destination is unknown and the waypoints are changing as the UAV moves. The result of the simulation shows that the flight path suited for monitoring and taking images is obtained by the method quickly.