The false space targets are mostly caused by stellar targets, smears of stellar targets and space targets, background image and image noise. The performance of eliminating them directly affects the detection probability and false alarm rate of the whole system. Based on the detailed analysis of the origin of false space targets, the methods of eliminating them are proposed according to their different features in images: 1) eliminating false space targets using their basic imaging features; 2) eliminating the false targets on smears of stellar targets and space targets by detecting the smears; 3) eliminating false space targets which come from stellar targets based on the changing of the temporal difference processing images and based on the variety of region images which include targets respectively; 4) confirming space targets with track information. Finally, combining with the hardware platform and real-measured data, the effectivity and real-time performance of methods are validated.