In order to effectively improve the fault detection and monitoring performance of the support vector machine (SVM) algorithm in the industrial process fault detection algorithm based on DW-ICA-SVM is proposed. Firstly, standardize the training data, used Independent Component Analysis (ICA) to obtain the independent component matrix of the data, extract hidden non-Gaussian information and generate IC. Then used the Durbin-Watson (DW) criterion to calculated the DW value of the IC. The important IC is extracted by comparing the DW value, and the training data containing important IC information is used as the input of the SVM model to obtain the weight vector and displacement. Finally, the test data is input to the model for fault detection and monitoring. While extracting hidden non-Gaussian information, this method uses DW method to effectively extract important noise information, reduces autocorrelation between samples, reduces non-random behavior, and combines SVM for fault detection, which effectively improves fault detection performance. The method is applied to nonlinear numerical examples and Tennessee-Eastman industrial process, and compared with PCA, LPP, ICA, SVM and ICA-SVM methods, the simulation experiment results further verify the effectiveness of the algorithm.