Aiming at the problem that the bearing fault characteristic signal is susceptible to noise interference, which leads to the difficulty of extracting the bearing fault impact characteristic signal. A bearing fault diagnosis method using the combination of the Chaos Sparrow Algorithm (CSSA) and the Maximum Correlation Kurtosis Deconvolution Algorithm (MCKD) is proposed. First, construct the adaptive function of CSSA based on the principle of kurtosis. Then, the CSSA algorithm is used to find the optimal period T and filter length L. Finally, the optimized MCKD algorithm is used to extract the faults of the motor bearings. And compared with unoptimized MCKD, particle swarm optimization optimization maximum correlation kurtosis deconvolution algorithm (PSO-MCKD), and Sparrow algorithm optimization maximum correlation kurtosis deconvolution algorithm (SSA-MCKD). The experimental results show that the CSSA algorithm has a faster convergence rate and better global search ability when searching for MCKD parameters compared to the particle swarm optimization (PSO) and the sparrow algorithm (SSA) algorithm. The proposed CSSA-MCKD method can effectively enhance the fault extraction ability of the MCKD algorithm, and has a faster convergence speed and global search ability.