Aiming at the system selection and design scheme of remote sensing stereo mapping satellites, combined with the current status of space mapping capability, this paper proposes a multi-baseline space mapping model by analyzing the advantages and bottlenecks of the traditional large, small and double baseline aspect ratio mapping models, to adapt to the mapping tasks of different terrain conditions such as mountain buildings. According to the theoretical basis for the derivations of factors affecting positioning accuracy under the stereoscopic imaging model, a numerical simulation scene is constructed, and the relationship between different matching errors and the plane and elevation accuracy is obtained by combining the analytical aerial triangulation method. The simulation results show that for different degrees of matching errors, the smaller ratio of baseline to altitude of about 0.15 to 0.6 is generally suitable for terrain characteristics with high requirements for the plane accuracy, and the larger larger ratio of baseline to altitude of about 0.8 to 1.2 is suitable for terrain characteristics with high requirements for the elevation accuracy. Accurate selection of the ratio of baseline to altitude according to the corresponding terrain can improve the positioning accuracy. The research results of this paper provide a reference for guiding the design of satellite mapping capability improvement schemes.