Abstract: In order to solve problems that conventional time-frequency domain analysis methods were sensitive to noise and had poor adaptability of energy change in coal-rock interface identification based on vibration signal analysis, a vibration signal identification method of coal-rock cutting of shearer based on cepstral distance was proposed. By analyzing cutting vibration signals of shearer acquired by vibration sensor under different load conditions, following conclusions are gotten. Compared with condition of cutting rock, the cepstral distance between vibration signal under condition of cutting coal and standard signal under no load condition is larger. The cepstral distance of the vibration signal under condition of rock cutting is obviously periodic, and the cycle is one rotation cycle of shearer drum, while the vibration signal under condition of cutting coal has no periodic feature. The industrial test result shows that the method has identification accuracy of 75% when hardness difference between coal and rock is more than 10 MPa.