Aiming at the problem that early fault features of rolling bearings are easily affected by strong background noise and difficult to be extracted, a method based on whale optimization algorithm was proposed to optimize the parameters of VMD(variational mode decomposition) and MCKD(maximum correlated kurtosis deconvolution) to diagnose rolling bearing faults. Firstly, the whale optimization algorithm was used and the minimum value of envelope entropy was selected as the objective function to optimize the number of decomposition layers K and the penalty factor α in the VMD algorithm to obtain the VMD decomposition results, and the optimal components were selected based on the weighted correlation crags for reconstruction. Secoxndly, the T-value of the MCKD algorithm was obtained based on the reconstructed components’ envelope spectra, and the whale optimization algorithm was used and the maximum value of the weighted envelope spectral crags was selected as the objective function. Optimize the filter length L and the number of displacements M in the MCKD algorithm, and enhance the shock component in the reconstructed component of the reconstructed signal through the MCKD algorithm. Finally, the type of faults of the rolling was determined bearing through the analysis of the envelope spectrum. The results of both simulation analysis and experimental tests show that the method has the ability to effectively extract and significantly enhance the shock component from the fault signal, and the multifrequency of the fault frequency can be clearly seen. Even in the case of strong background noise interference, the method is still able to successfully realize the accurate diagnosis of rolling bearing early faults.