Aiming at the problem of effective extraction and identification of rolling bearing fault information in complex environments, a fault diagnosis method for rolling bearings based on feature mode decomposition (FMD) combined with multiscale fuzzy dispersion entropy (MFDE) and zebra optimization algorithm (ZOA) optimization support vector machine was proposed. In order to solve the problem that the key parameters in FMD are not adaptive, the minimum envelope entropy was used as the objective function, and the beluga whale optimization (BWO) was used to optimize FMD to find the optimal parameter combination to achieve the optimal decomposition of fault signals. Multiscale fuzzy dispersion entropy was introduced to construct the eigenvectors under different modes after decomposition. Finally, the feature vectors were input into the support vector machine for training and recognition. The effectiveness of the proposed method was verified by the public dataset and the self-made experimental platform dataset.