This paper presents a 3Dimensional-GNSS (3D-GNSS) positioning technique, which is used for localization and mapping simultaneously (SLAM). The 3D building model becomes an important aid to many positioning techniques such as LiDAR and GPS positioning methods. In order to automatically create the accurate map in wide area, the precise position of mobile mapping platform is needed. However, GNSS positioning performance is severely degraded because of the effects of multipath and Non-Line-Of-Sight (NLOS) in the urban area. With the aid of 3D building map, the proposed 3D-GNSS distinguishes whether the received GNSS signal is transmitted as LOS or NLOS path, and can calculate the length of reflection path for the improvement of positioning error. To achieve highly accurate positioning in urban canyon, we further develop 3D-GNSS based integrated vehicle self-localization system, which is comprised of 3D-GNSS, inertial sensor and vision sensor. On the other hand, highly precise positioning needs the accurate 3D building map. Inspired by the idea of 3D-GNSS, this paper proposes to optimize the 3D building map by estimating the reflection of GNSS signals. In addition, the position of mobile mapping platform is estimated from our integrated localization system in the mapping stage. The experimental result demonstrates that sub-meter accuracy is achieved in both localization and mapping.