Information security is one of the cornerstones of Information Society. Integrity and privacy of financial transactions, personal information and critical infrastructure data, all depend on the availability of strong and trustworthy security mechanisms. In recent years, many researchers are using data mining techniques for building IDS. Here, we propose a new approach by utilizing data mining techniques such as neuro-fuzzy and radial basis support vector machine (SVM) for helping IDS to attain higher detection rate. The proposed technique has four major steps: primarily, k-means clustering is used to generate different training subsets. Then, based on the obtained training subsets, different neuro-fuzzy models are trained. Subsequently, a vector for SVM classification is formed and in the end, classification using radial SVM is performed to detect intrusion has happened or not. To illustrate the applicability and capability of the new approach, the results of experiments on KDD CUP 1999 dataset is demonstrated. Experimental results shows that our proposed new approach do better than Conditional random fields (CRF) with respect to specificity and detection accuracy.