A ball screw pair fault diagnosis method based on sparrow search algorithm optimized variational modal decomposition algorithm (SSA-VMD) combined with support vector machine (SVM) was proposed for the problem of difficult extraction of ball screw pair fault features.The autonomous optimization of the VMD parameters was performed using the minimum envelope entropy as the fitness function of the SSA.The IMF energy value was used to filter and reconstruct the decomposed signal to remove the noise and irrelevant components.Finally,8 types of time-domain feature parameters and 5 types of frequency-domain feature parameters of the reconstructed signal were extracted as the feature vector set,which were imported into SVM for training the fault recognition model.Vibration signals were collected by building a ball screw pair fault diagnosis experiment platform.The SSA-VMD,VMD,and EMD methods were used to decompose the signals and extract the fault features,respectively.The experimental results show that compared with VMD and EMD,SSA-VMD can autonomously select the optimal VMD parameters for signal decomposition with different signals,and can accurately identify the type of ball screw pair fault.The results demonstrate the feasibility and accuracy of SSA-VMD-based ball screw pair fault diagnosis.