In recent years, the research of 3D mapping technique in urban environments obtained bymobile robots equipped with multiple sensors for recognizing the robot’s surroundings is beingstudied actively. However, the map generated by simple integration of multiple sensors data onlygives spatial information to robots. To get a semantic knowledge to help an autonomous mobile robotfrom the map, the robot has to convert low-level map representations to higher-level ones containingsemantic knowledge of a scene. Given a 3D point cloud of an urban scene, this research proposes amethod to recognize the objects effectively using 3D graph model for autonomous mobile robots. Theproposed method is decomposed into three steps: sequential range data acquisition, normal vectorestimation and incremental graph-based segmentation. This method guarantees the both real-timeperformance and accuracy of recognizing the objects in real urban environments. Also, it can provideplentiful data for classifying the objects. To evaluate a performance of proposed method, computationtime and recognition rate of objects are analyzed. Experimental results show that the proposedmethod has efficiently in understanding the semantic knowledge of an urban environment.