Abstract: With the rapid development of the manufacturing industry,industrial robots have become the core execution unit of intelligent manufacturing.Aiming at the problem that the input speed of rotary vector（RV） reducer is time varying and not easy to be read during operation,a health state evaluation method of RV reducer is proposed with the combination of current signal and vibration signal.Based on the instantaneous frequency of the current signal,the input rotation frequency information is obtained.The fault characteristic frequency of the reducer components is calculated,and the stationary vibration signal is intercepted synchronically.The extracted vibration characteristics are decomposed using the variational mode decomposition algorithm of particle optimization,and the fault sensitive component is selected based on the multidimensional evaluation function for envelope analysis.Fault information is mined to evaluate the health status of the fault,and the effectiveness of the method is verified by the experimental data set.The results show that the proposed method can obtain the joint input speed of the industrial robot during operation.The multi-dimensional evaluation function can extract the fault sensitive component rich in shock information from the RV reducer signal.