3D mapping for unknown environments is a crucial issue for mobile service robots, which has many important applications in navigation and path planning. In this paper, we use Microsoft kinect sensor mounted on a mobile robot to obtain 3D map. During the process, consistent frames of RGB-D data are captured and integrated. Iterative Closest Point (ICP) algorithm and incremental registration method are employed to perform the fine registration. Furthermore, we use OctoMap to store 3D map of octree representation, an efficient data structure. Finally, our results demonstrate the practical applicability and effectiveness of the proposed approach.