IDS (Intrusion Detection System) is used to monitor N/W traffic or activity or file modification. If the system is being targeted by unauthorized person or intruder such as denial of service attack, SYN attack etc., it can be detected by the IDS. But Many times the system gets vulnerable to the new attacks. Today, most of Intrusion detection system cannot identify new attacks in the network for incoming packets from internet. So proposed model detects these unknown attacks with the help of optimized decision tree from available set of datasets. The proposed model presents an approach for building an intrusion detection system for a network by using hybrid classification model.