This paper presents three dimensional (3D) map building method for the intelligent vehicles based on accurate indoor localization using a mobile mapping system (MMS) that is equipped with perception sensors consist of a wheel odometer, a laser range finder (LRF), and two projected texture stereo (PTS) cameras. The environmental data measured by perception sensors are stored in the node units according to a certain distance interval. In order to estimate the positions of the MMS using the relationship among nodes, the localization method is divided into two parts, front-end (map-based scan matching) and back-end (graph-based optimization). The estimated positions are used to build the grid-based map and the point cloud dataset, respectively as the 2D and the 3D maps through the mapping process (Bayesian model). An experiment has been performed in office environment (indoor) to verify the effectiveness of the proposed method. Experimental results show the high precision of 3D point cloud dataset that can be used for various applications including navigation of intelligent vehicles and pedestrians in indoor environments.