Information and communication technology inflate day by day, due to rapid improvement in technologies has increased the need of effective IDS (Intrusion Detection System). Here, Intelligent Intrusion Detection method that is Rough Set based approach presented for performance evaluation of classifier abnormal behavior. Rough Set Theory is used to reduce the input data space, from complex databases and find minimal decision rules or reduct, through this we can manage complexity of system and manage huge network traffic. Rough set based effective classification models namely Rule based classifier algorithm with discretization, Decomposition tree algorithm and Decomposition tree with discretization have been applied to find reduced decision rules and classify problem. Comparison of classification results also have perform with various evaluation criteria and recognize best suited classifier for intrusion detection system dataset. This paper tries to achieve better accuracy and coverage using rough set based classifications. Then compare classification models performance using ROC(Receiver operating characteristic) curve with optimal visualization of results. Comparison of Classification models performance using ROC also done. These results provide robust, high accuracy, lowest false alarm rate and improve coverage of test data set. An empirical result shows the optimum classification result to big network data traffic.