Abstract: Tracking model is one of main models, when star sensor works. Speediness of tracking process affects star sensor′s performance. A fast star tracking algorithm was proposed. Three methods, such as star catalog partition, threshold mapping, sorting and matching, were designed for three time-consuming elements in tracking algorithm, in order to raise the tracking speed. Star catalog partition method divided the celestial sphere into some partitions. It made star mapping search in the partitions near the direction of star sensor′s visual axis not in the whole celestial sphere and reduced the total number of stars searched. Threshold mapping method in required accuracy condition, set a threshold of stars tracked. When the number of stars tracked was less than the threshold, star mapping was performed. It reduced mapping times. Sorting and matching method made sorting based on star′s coordinates in star image, and then matched. It avoided matching between stars which had a long distance. Simulation showed that application of these methods made this tracking algorithm faster, and improved the efficiency of star sensor.