Ultra-wide band (UWB) micro-wave imaging has been proposed as a new method to diagnose breast tumors. An artifact is considered one of the imaging issues. This study focuses on current methods in this field. First, different artifact removal approaches are reviewed, where the aims of most signal processing methods include skin artifact removal and diagnosis of the existence and position of malignant lesions. Then, to improve the artifact removal, we design a signal with linear frequency modulation (LFM) intra-pulse modulation and constant pulse repetition interval (PRI) inter-pulse modulation. Here, the received and submitted signs of the skin are removed by using the auto-correlation function. The simulation results show that this proposed method has minor errors compared to the other methods.