With the development of new energy grid connection and ultra-high voltage (UHV) DC transmission, the requirement of reactive power regulation for power grid has gradually increased, so large synchronous condensers have been put into use again. A fault diagnosis method based on stochastic subspace identification (SSI) and multi-core support vector machine (MSVM) is proposed to facilitate the fault diagnosis of the synchronous condenser bearing. The vibration sensors are used to collect vibration signals at different positions on the outer surface of the synchronous condenser bearing, and the random subspace model is used for feature extraction. According to Gaussian support vector machine (SVM) and multi-core learning method, a multi-core SVM is constructed. Then, the extracted feature data are imported into the MSVM for fault diagnosis. The experimental results prove that the synchronous condenser bearing fault diagnosis method based on SSI-MSVM is suitable, and the fault can be successfully identified.