The images produced by the CMOS sensor of the high-orbit satellite′s optical axis pointing measurement system are prone to degradation due to the harsh space environment. This degradation adversely affects the quality of the images and compromises the accuracy of the spot centroid results. To combat this issue, this study proposes a method for high-precision, real-time determination of spot centroid. Simulation analysis results show that this method can accurately track changes in the spot position within images, validating the feasibility of its step-by-step calculation strategy. The study employed the progressive method, OTSU centroid method, and the Gaussian fitting method for the real-time computation of spot centroids from images acquired by the measurement system. The experimental results show that the progressive method has good real-time performance, the highest success rate of spot location, and the smallest average deviation and standard deviation of centroid calculation, being 0. 026 and 0. 029 pixels, respectively. This approach provides reliable data support for optical axis pointing measurement systems.