The factors, which influence the distribution of the amplitude and pha se of the points in the phase interferogram of SAR-ATI moving targets indicatio n , including radar-noise, scatter energy, background uneven RCS and the side lob e of image pixels were analyzed. To resolve the problem of the high fal se alarm rate in SAR-ATI, an improved method was proposed based on the MTI of th e joint amplitude-phase threshold. Firstly, the points with polluted phase in th e image of every channel were pre-filtered. Secondly, after eliminating the point s with invalidated phase, the adaptive amplitude filtering was processed with the threshold set by the probability, according to the radar-noise and the phase in t erferogram. In the end, the detection with the phase threshold method was perform ed to reduce the false alarm rate effectively. An empiric al formula derived from the statistic fitted curves in the phase interferogram was presented to reduce the false alarm rate in the low CNR condition evidently. The computer simulation demonstrates this method.