To monitor the operation state of electric motor by applying the noise of motor, and judge whether the operation state of the motor is abnormal and the cause of the abnormality by combining the monitoring results. The concept of spectral centroid factor is proposed considering the mechanism and characteristics of motor noise, and the spectral centroid analysis model is constructed. The paper studies the basic principle and implementation approach of the spectral centroid factor, and adopts software to conduct the simulation analysis. The analysis results show that the parameter is capable of clearly extracting the fine features of motor noise, reflecting the operation state of the motor in real time, and can be referred as an effective basis for motor operation state monitoring and failure analysis. Based on this, the paper proposes a new monitoring approach for motor operation status, and conducted experimental verification, which is of strong theoretical research value and significantly practical application value.