Abstract: In the field of coal mine production, the mine hoist, as a kind of auxiliary transportation equipment, plays a very important role in the whole transportation engineering of the mine. Its safety and stability directly affect the production efficiency of the coal mine and the safety of underground workers. After the failure of mine hoist, its sound signal will also change with the operation state of the equipment, so the abnormal operation state of the equipment can be detected by analyzing the characteristics of the sound. In order to ensure its safe and reliable operation and improve coal mine transportation efficiency, an accurate fault identification method of mine hoist based on MFC-CS-MUSIC was proposed. By collecting the hoist audio signals, MFCC algorithm was used to extract the Cepstrum coefficients of multiple channels for fault identification. The acoustic signal after MUSIC fault identification is used to locate the signal and the minimal direction of arrival is obtained. The DOA value obtained by MUSIC algorithm is taken as the optimization variable, the difference between calculating DOA and measuring DOA is taken as the objective function, and the CS algorithm is used to optimize the objective function, so as to realize the precise location of the hoist fault source. The experimental and application results show that the azimuth error Δψbetween the music algorithm and the actual position coordinate after CS algorithm optimization is less than 5°, Δθis less than 4°. This method can accurately identify the hoist fault and accurately locate the hoist fault source, greatly shorten the troubleshooting time of the mine hoist fault location, and significantly improve the work efficiency of the mine hoist.