Abstract: In order to realize automatic correction for smear effect of high light target in space observation image, an automatic de-smear system was established, and its applied algorithms such as Gaussian noise distribution parameters estimation of image background, target smear detection and image gray-scale value correction were investigated. First, the Gaussian noise-s distribution parameter of star observation image was estimated with histogram least square curve fitting. Subsequently, utilizing the smear features in observation image, a smear detection algorithm based on statistical information was proposed. Then, after the smear position was determined, the contaminated pixel-s gray-scale value was corrected. Finally, a set of de-smearing system in star observation image was with Visual Studio 2005. Experimental results indicate that with the star observation image of 16 bit and 1 0241 024 pixel, the processing time of single frame image is about 300 ms. The smear effect is well corrected, and the useful information of stars and target is not destroyed. The processed image by our algorithm can satisfy the demands of stability, reliability and precision for dim target detection in space observation image.