Abstract: A seismic-while-mining(SWM) monitoring system laid in a working face of an underground coalmine will continuously generate tremendous amount of SWM data. When a coalmine cutting machine cuts coal seam actively, the SWM data will be of high correlation, and when it is not, the SWM data will be of low correlation or even no correlation at all. Based on this characteristic, a method for automatically and quantitatively evaluating the quality of SWM data was proposed. Using this method to process SWM data from a coal mine in Inner Mongolia, the result showed that this method can recognize the correlation peak very effectively, and the data quality evaluation result is very good. To apply this method to screening SWM monitoring data from a coal mine working face in Guizhou Province, the stack result of selected high quality data shows that the S/N ratio is obviously improved, and the correlation event is obviously enhanced. This method can be used to rapidly pick out high quality data from enormous amount of data, and reduce the workload of next step while improving the processing result.