Aiming at the problem that there is a large amount of noise in the contactor characteristic parameter data, which interferes with the construction of the contactor performance degradation index, this paper completes the noise reduction of the contactor characteristic parameter based on the relative entropy optimisation Variational mode decomposition (VMD) and Non-local mean (NLM) noise reduction, and reasonably constructs the contactor performance degradation index based on the convolutional autoencoder (CAE). Firstly, the voltage and current signals are collected based on the full-life test platform of the contactor and the characteristic parameters of the contactor are extracted; Then, the parameters of the VMD are optimised by relative entropy and the dominant and effective components of the noise are selected based on the correlation; Secondly, the dominant signal of the noise is denoised by the NLM, and the noise-reduced signal is obtained by signal reconstruction. Finally, the contactor performance degradation index is constructed by CAE. The final results show that the method of this paper can better complete the contactor characteristic parameter noise reduction and performance degradation index construction.