Abstract: Coal and rock identification is the key technology to realize an automatic adjustment of shearer drum height. Reliable coal and rock identification system has outstanding advantages in improving production efficiency,reducing e- quipment wear,and ensuring miners’ safety. At present,the coal-rock interface is mainly identified by miners or sin- gle-sensor monitoring,so the recognition results are inaccurate and have some errors. Therefore,a multi-sensor coal- rock recognition method based on data fusion theory is proposed. This method is based on the difference of hardness of coal and rock,and uses the pin shaft of shearer rocker arm for data collection. Four pin shaft sensors with equivalent strength treatment are arranged between the rocker arm and the connector,and the strain data of the pin shaft when the shearer cuts different hardness coal wall and rock wall are collected,and the collected data are distributed and merged by weighted fusion theory. The combination criterion of multi-sensor data fusion is obtained,and the coal-rock interface is identified by the combination criterion. The experimental results show that the strain data collected by pin shaft sen- sor fluctuate greatly in coal and rock cutting,and there is overlap between coal cutting strain data and rock cutting strain data,so it is difficult to accurately identify coal and rock interface. The fluctuation of strain data processed by data fusion method is small. The load value of the pin shaft sensor is obtained by calibrating the strain data with the fit- ting formula:the stress range of the pin shaft is 24. 766-25. 467 kN in cutting rock and 23. 493-24. 348 kN in coal cutting. Compared with the single sensor measurement,the difference is obvious,and there is no overlap part. There- fore,this method can be used to calibrate the expected value range of cutting coal and rock pin shaft force in practical production,and the difference of expected value range can be used as the basis for the coal and rock interface identifi- cation of shearer.