Abstract: In order to improve the automation and intelligence of top coal caving, and lay a foundation to realize the target of less quantities of workers even no wor ker in the fully- mechanized top coal caving face, using the self- developed mine portable vibration data recorder, the time- domain indexes of vibration signals at the hy draulic support tail beam and the rear scraper conveyor were acquired under the different conditions of coal caving, gangue caving and roof rock caving in the fully- me chanized top coal caving face. The above signals were extracted and comparatively analyzed, some conclusions were drawn that the variance, skewness and kurtosis i ndex of the vibration signals were sensitive to working condition changes, and the kurtosis index was the most sensitive. These indexes could be used as the evaluation indicator for the coal and rock character recognition in the fully- mechanized top coal caving face. Due to the complexity of the vibration signal acquisition, the measure pointed at the rear tail beam of sublevel caving hydraulic support was more suitable than the rear scraper conveyor for the coal and rock trait identification in the fully- mechanized top coal caving face.