Threat from an eavesdropper has always been a concern for government and other private organizations. Be it in the form of traffic analysis, modification of messages, disruption of normal network services, or identity mismatch, attackers have always come up with ways to disrupt normal communication between two parties. They try to intrude into a server or a system with malign motives which eventually harms the organizations and government alike. Intrusion Detection System is a very naïve field which focuses on the detection of abnormal network traffic and reports in real time if any malicious activity is detected. The main objective of our proposed work is to detect the intrusions and identifications of security threats and reduce false alarm rate. In our current research work, Spider Monkey Optimization (SMO) algorithm is used with Artificial Neural Network (ANN) to detect attacks or intrusion in the system. The developed model was rigorously examined on a publicly available dataset LUFlow20 to classify any incoming traffic as attack or normal. The ANN-SMO model gives an accuracy of 99.82% on LUFlow20, which is a labelled dataset.