Slewing bearing is the key component of wind turbine and is used to transmit radial and axial load as well as the resulting overturning moments. The poor working condition will easily result in fatigue failure. An effective method for predicting the residual useful life of slewing bearing is proposed. Firstly, the features of temperature, torque and vibration signal of service sample and reference sample are extracted separately. Second, principal component analysis (PCA) based multiple sensitive features is used to establish performance decline indicator. Further analysis on these three PCA indicators is made by Support Vector Data Description (SVDD). Then the similarity is calculated between service sample and reference sample by normalized cross correlation (NCC) and residual useful life of service sample is predicted according to the life of reference sample. Finally, the method is verified by two experiments based on different working conditions. The prediction absolute error is only 0.9% when interval length is 50.