A novel fault identification method based on Supervised Incremental Locally Linear Embedding (SILLE) is proposed in this paper. The time-frequency domain feature set is first constructed to completely characterize the property of each fault. Then, SILLE is introduced to automatically compress the high-dimensional time-frequency domain feature sets of training and test samples into the low-dimensional eigenvectors which have better discrimination. Finally, the low-dimensional eigenvectors of training and test samples are input into Morlet wavelet support vector machine (MWSVM) to carry out fault identification. SILLE considers both local manifold geometry and class labels in designing the reconstruction weight matrix and applies local linear projection to obtain the embedded mapping of the new fault samples, thus, it improves the fault identification accuracy and achieves rapid incremental processing of the new samples. Fault diagnosis example on deep groove ball bearings and life state identification example on one type of space bearing demonstrated the effectivity of proposed fault identification method.