Abstract: In this paper， a target localization algorithm based on improved Hough transform frequency fitting technology is proposed. By dynamically adjusting two shape parameters， the adaptive extended Bezier frequency fitting model is constructed. The echo signal is demodulated by the fitting curve to complete the separation of multiple target components， and the target motion trajectory is synthesized by combining Doppler processing method to realize the real-time target positioning and tracking. In order to solve the problem of large computation of multi-dimensional parameter search in the process of fitting model construction， according to the relatively independent characteristics of two parameters， linear search is used to replace the traditional multi-dimensional search， which greatly reduces the computational complexity and optimizes the real-time performance of target detection. The simulation and experimental results show that the algorithm can effectively suppress the frequency ambiguity in the multi-body target tracking and positioning scene， and further improve the applicability when the target trajectory is complex and changeable.