Abstract: With the development of 5G and even 6G wireless communication technology in the future, the number of wireless communication devices presents an explosive growth trend.In contrast, the electromagnetic spectrum environment is increasingly congested, and the traditional communication spectrum that is nearly exhausted can no longer meet the surging business demand. In this context, the integrated signal of spectrum sharing for radar and communication has attracted great attention from the industry and academia. However, in the framework of matched filtering, the integrated signal couldn’t take into account the performance of radar and communication.The communication information was bound to produce high sidelobe and false peak in the radar ambiguity function. For this reason, some scholars proposed a misfit processing method based on Orthogonal Frequency Division Multiplexing (OFDM) shared signal, which extrapolated the high sidelobe and pseudo peak to the outside of the radar observation window, in order to give consideration to the radar fuzzy performance. However, this method had produced SNR loss, and the SNR loss increased with the increase of observation window. In view of this, this paper proposes a fusion mismatch processing and Least Mean Square (LMS) filtering algorithm. Through the deep fusion of LMS and mismatch processing, the constraint between SNR loss and observation window width can be broken, and then the SNR loss can be reduced without reducing the observation range, or the observation range can be greatly increased with the same SNR loss.