Abstract Aiming at the problems in the defense of near-space high-speed vehicle, this paper studies the motion state estimation algorithm of this kind of vehicle. First, the target maneuvering model is analyzed, then the radar observation model is established, and the fusion algorithm of SRCDKF based on interaction multiple model (IMM) is studied. Finally, based on the typical jumping reentry trajectory of near-space high-speed vehicle, the target tracking error curves of different filtering algorithm are given through simulation, and the effectiveness of IMMSRCDKF filtering algorithm for near-space high-speed vehicle information estimation is proved.