Abstract: In order to improve the accuracy of fault diagnosis of rolling bearings,this paper proposes a fault diagnosis method for rolling bearings of rotating machinery,which based on variational mode decomposition（VMD）and scalable and mutational particle swarm optimization（SVPSO）to optimize BP neural network. By introducing scaling factors and particle mutation operations to improve the local and global optimization performance of the standard particle swarm algorithm,an improved particle swarm algorithm-Scalable and Mutational particle swarm algorithm（SVPSO）is obtained,and then the algorithm is used to optimize the values of the weights and thresholds of the BP network to improve the fault diagnosis accuracy of BP neural network. Furthermore,for reducing the impact of the input feature vector on the classification performance of the BP neural network,VMD is used to decompose the bearing vibration signal and calculate the time-frequency entropy of its IMF component to construct the signal feature vector. By comparing with other diagnostic methods using the same benchmark bearing data set,the fault diagnosis accuracy and algorithm stability of the method proposed in this paper has been effectively improved.