Today, world has come closer due to rapid increase of internet. So computer security is of big concern now. As technology has been developed many threats are emerged for the data security which is not at all good for sensitive data transactions. So it is necessary to build high level security to provide safe communication between various networks. Intrusion detection systems are built to detect the attacks. Because of the intruders, the security of the network has become serious problem. Thus to overcome this we are proposing this paper for intrusion detection using k-means and Naïve Bayesian Classifier, which is depend on probabilistic model. This algorithm performs attack detection and keeps false positive rate at low level for different types of networking attacks