With the approaching era of the web, network security has become the key foundation for a ton of economic and business net applications. Incursion detection is one of the looms to resolve the matter of network security. The imperfection of incursion detection systems (IDS) has given a chance for data processing to make many vital contributions to the sphere of incursion detection. In recent years, several researchers are mistreatment data processing techniques for building IDS. Here, we propose a brand new approach by utilizing data processing techniques like neuro-fuzzy and radial basis support vector machine (SVM) for serving IDS to achieve a higher detection rate. The projected technique has four major steps: primarily, the k-means bunch is employed to get totally different coaching subsets. Then, supported the obtained coaching subsets, totally different neuro-fuzzy models are trained. Later, a vector for SVM classification is made and within the finish, classification mistreatment radial SVM is performed to notice incursion went on or not. Maybe the applicability and capability of the new approach, the results of experiments on KDD CUP 1999 dataset is incontestable. Experimental results show that our proposed new approach does higher than BPNN, multiclass SVM and different well-known strategies like call trees and Columbia model in terms of sensitivity, specificity and specifically detection accuracy.