Abstract: Different from the LFMCW signal, the radar adopting OFDM system signals has high sidelobes in its ambiguity function when the transmission information modulated on each symbol is completely random. When there are many strong clutters in the detection scene, the range-Doppler peak corresponding to the target echo will be submerged under the sidelobes of the strong clutter after pulse compression accumulation, which will lead to the inability to identify the target. Currently, there are many clutter rejection schemes to deal with this problem, such as extensive cancellation algorithm and methods based on compressed sensing, but these methods often face the problem of insufficient clutter suppression. Based on this, we propose an efficient processing scheme based on greedy strategy, which can suppress strong clutter point by point adaptively to reduce the sidelobe base and make the target information emerge. We have compared it with the measured data, and this method achieves better clutter suppression performance than conventional methods.