This paper presents a combined discrete wavelet transform (DWT) and supportvector machine (SVM) technique for fault diagnosis of spherical roller bearing application in high performance hydraulic injection molding machine.The proposed technique consists of preprocessing the fault vibration signal samples using discrete wavelet transform at the fourth level of decomposition of fault vibration signal for fault classification.After feature extraction from fault vibration signal,support vector machine is used for decision of fault types of the spherical roller bearing.The classification results indicate the effectiveness of the combined DWT and SVM based technique for fault diagnosis.