Rolling bearing is one of the key components of rotating mechanical equipment, which performance degradation assessment is the basis of condition monitoring and conditionbased maintenance of mechanical equipment.A method of evaluating the performance degradation of rolling bearing based on single layer sparse autoencoder(SAE) and support vector machine(SVM) was proposed in order to accurately grasp the degradation trend and the degree of performance of rolling bearing.Sparse autoencoder learning method by which the potential implicit information of data could be deeply exploited and the extraction method of bearing state feature based on timefrequency domain and sparse autoencoder were studied.An improved evaluation algorithm of bearing performance degradation based on SVM classification algorithm was proposed and applied to the performance degradation evaluation model of rolling bearing. The method of model parameter optimization was determined. Finally, the performance degradation trend of the obtained feature input performance degradation evaluation model was obtained and an example of rolling bearing was given to demonstrate the practicability of the proposed method.