Abstract: In view of the single event upset (SEU) that may occur when a large number of low-cost and high-integration components are used in micro star sensors on commercial spacecraft, methods of monitoring the star tracking process, program comparison and correction, watchdog timeout, and reset type detection are proposed in this article. By the software approach, SEUs can be autonomously restored, while sending reset related information to the attitude & orbit control subsystem, to assist the subsystem in determining whether to intervene in the star sensor based on the product status. Ground testing verification proved that faults caused by SEUs can be effectively recovered. The time from fault occurrence to fault discovery was less than 19 s, while the time from fault discovery to restoration to normal attitude output was only 0.2 s, which was one thousandth of the time required for ground judgment and control. The above anti-SEU scheme may provide guarantee for the auto-recovery from SEUs occurring in the micro star sensor on spacecraft in cases where satellite telemetry and remote control cannot be performed by ground base stations.