In order to shorten the revisit period, and improve the efficiency of imaging, currently, all the in orbit high resolution remote sensing satellites adopt side-view imaging technology. When satellites imaging with a side perspective, it would inevitably cause the degradation of image quality, like spatial resolution reduction, image deformation incensement, and the reduction of positioning accuracy. In addition, with the increase of the side perspective, the error of DEM data will be amplified in the process of ortho-rectification, which, thus, will bring bigger error for image mapping accuracy. To address this problem, embarked from the impact of side view and DEM precision on image point error, this paper used Quick bird image data of the same area but different side perspectives, and simulate the impacts of different DEM elevation error on different side view image. At last we put forward a method to determine the maximum side perspective with map precision constraints, and set up a conversion relationship among satellite side view, DEM error and image point error. Finally, the paper combines the errors of 1:50000 DEM, ASTERGDEM and the requirement of ortho-image drawing, discusses and gives the values of maximum side perspective in the mountains and plains.