Abstract: Seismics while mining is a passive seismic detection technique using a shearer as seismic source. Because the shearer moves unceasingly and excites the continuous signals, the conventional data processing method cannot be applied directly. In the paper a cross-correlation method of segmented waveform was put forward, the effective signal travel time was extracted through the segmentation of the data generated the shear, and then the velocity imaging method was used to image the interior of the working face and the front of the cutting hole. The wave equation was used to carry out the seismics while mining numerical simulation of the working face containing a abnormal body. The imaging results from the inversion were basically consistent with the velocity model. In the actual seismics while mining data test, this method is used to get the imaging results of the seismic wave propagation velocity obtained from the inversion of the data of the shear source during mining, so as to realize the real-time dynamic monitoring of an area of the abnormal variation in the working face. The results indicate that the cross-correlation imaging method based on the segmentation can solve the problems in the data processing of the signals of the shear source, meet the requirements of real-time and stability of seismics while mining technology.