Securing the network systems from intrusions is the main concern in network management because even a single intrusion within the network can cause heavy loss to the reliability and efficiency of the computer network. In order to prevent these attacks, detection is the first step for which intrusion detection systems (IDS) are designed. For these IDS, it is important to be trained and tested with all possible types of attack in networks. For this purpose Knowledge Discovery and Data mining (KDD Cup) is the standard datasets containing variety of attacks. Accuracy of the IDS depends upon the appropriate classifier selection which can classify the intrusions successfully from the normal data. In this paper different classifiers are presented and also their performance is evaluated on standard dataset.-This will assist the researchers in determining optimal classifier for their IDS depending upon their effectiveness and efficiency in detection of attacks. © 2013 International Information Institute.