A high-efficiency dim space object recognition method, the steps are: 1) star image imaging, to obtain a series of star images containing faint objects; 2) star image processing, to extract blocky bright spots. 3) Stellar removal, which mainly removes most of the star elements from all bright spots; 4) Obtains the motion characteristics of each blocky bright spot in the suspected target queue; 5) Determines the ordered feature set; 6) Target tracking, for the established Confirm the target queue to track, predict the position of the target in the next frame, calculate the target parameters successfully, extract the precise orientation information, and add it to the confirmed target queue. The method of the present invention can remove interference elements, such as stars and noise points, etc. from the extracted bright spots to the greatest extent, and finally realize tasks such as detection, extraction, identification, and tracking of space dim targets, and is especially suitable for low signal-to-noise ratios, complex star map backgrounds, multiple Various occasions where target information is unknown.