Recognizing its surroundings is important for biped robots seeking a destination. In this paper, we propose a motion planning method of a biped robot including path planning and obstacle avoidance. The robot obtains distance information on its surrounding environment from images captured by a stereo camera system, and generates a 3D map, then, builds a 2D grid map that locates flat floor regions, obstacle regions, bump regions, gate regions, and un-measured regions to decide its path by using the 2D grid map. Experimental results confirm the effectiveness of the proposed method.