In this paper, fault detection and diagnosis model using SVM (support vector machine) is proposed. This model is applied to develop inferential model that detects faults of a chemical process and diagnoses them. The existing multivariate method including PLS and Neural Network has been shown to be a powerful technique for process modelling and multivariate statistical process control. However, for systems that exhibit non-linear behavior and have sparse data, they can be inappropriate. In this paper by using SVM regression, more accurate and fast fault detection model is proposed. And then SVM classification is used for fault diagnosis model using known fault data. To verify the superior performance of the proposed model, the data sets of Tennessee Eastman Process are applied.