Abstract: In order to improve the intelligence of fastener disease diagnosis， a fastener condition diagnosis method is proposed based on vehicle dynamic response data and generalized demodulation time-frequency analysis combined with sparrow search algorithm-support vector machine （SSA-SVM） model. The acceleration signals of the normal and abnormal sections of the fastener are collected， and the short-time Fourier transform and the maximum overlapping discrete wavelet packet transform are used to preprocess the signal data. The generalized demodulation time-frequency analysis method is used to decompose the signal， and the effective value， energy contribution rate and wavelength of the main information components are calculated as the characteristic index. The characteristic index is trained by the joint SSA-SVM model to construct the classification model. The results show that the accuracy of the method is 97.50%， and several evaluation indicators are used to verify that its effectiveness and accuracy can meet the actual needs.