Aiming at the problem that the fault feature information of rolling bearing fault vibration signals is not fully extracted at a single scale,which results in low fault diagnosis recognition rate,a bearing fault diagnosis method was proposed based on granular dispersion entropy (FIG-DE) and SSA-SVM.The fuzzy information granulate was used to granulate the bearing vibration signal,and the fuzzy information granulation in three scales of fLow,fR and fUp was obtained; the dispersion entropy was calculated for the three groups of signals; the obtained entropy was composed of eigenvector matrix and input into SSA-SVM to carry out the bearing fault classification.The results show that the accuracy of SSA-SVM is significantly improved.