In order to improve the precision of the simulation system in a star sensor,this paper presents a gray pervasion method based on a Gaussian law for precisely simulating star spots and a Gaussian centroid extraction algorithm.The gray vaiues of simulating star image pixels are evaluated according to the 2D Gaussian distribution law and the mapping coordinates instead of the central pixel round coordinates are taken as the symmetric center,so as gray pervasion phenomenon caused by defocusing and aberration from an actual star spot can be simulated precisely.The new Gaussian centroid extracting sub-pixel positioning method consists of two steps.One is rough pixel-level positioning and the other is fine deviation positioning.A piecewise defined synthesized function is constituted based on Gaussian law to realize the sub-pixel fine deviation positioning.Under the simulating condition of image Gaussian noise in N(0,1.22) and the whole deviation interval of [-0.5,0.5) pixel,the Gaussian centroid positioning precision is 0.007 pixel,which is far less than 0.041 pixel from the gray-weighted centroid method and 0.026 pixel from the square-gray weighted centroid method.After disposing the simulation star map,the conclusions show that the Gaussian gray pervasion method for simulation star spots is rational and accurate,and Gaussian centroid extracting method has a higher precision than those of traditional gray-weighted centroid methods.