To address the bearing fault diagnosis problem, this paper proposes a multi-strategy fusion improved bagger optimization algorithm (ITDO) to optimize the variational modal decomposition-support vector machine (VMD-SVM) for bearing fault diagnosis. ITDO adds the DE-Rand-2 strategy, the period variation strategy, and the new move strategy on the basis of the TDO to form a new optimization algorithm, which is able to quickly find the global optimal solution and avoid falling into the local optimal solution, and then avoid falling into the global optimal solution. Quickly find the global optimal solution and avoid falling into local optimization. First, the improved bag badger optimization algorithm (ITDO) is used to optimize the modal components and penalty factor of the variational modal decomposition (VMD) to improve the decomposition accuracy. Then, the processed feature data are classified using SVM to identify different bearing fault types and normal states. In order to verify the effectiveness of the model on bearing fault diagnosis, the bearing fault dataset from Case Western Reserve University was used to classify inner ring, outer ring and rolling element faults. The experimental results show that the ITDO optimized VMD-SVM has the highest correct rate of bearing fault diagnosis compared with MSO, GWO and TDO optimized VMD-SVM.