Firewalls, cryptographic methods, and antivirus scanners are becoming increasingly ineffective in the face of increasingly complex threats. To make network channels and machines more secure, stronger protective walls are required. In addition to the policies already in place, an intrusion detection system can serve as an extra line of defence. This research suggests using a Naïve Bayes classifier in conjunction with an Enhanced Grey Wolf Optimizer Cuckoo Search Optimization (EGWCSO) to enhance performance. Min-Max normalization and 1-N encoding are used in the preprocessing step. EGWCSO is applied for the best feature selection, and Naïve Bayes is selected for classification. Single connection vectors, around 4,900,000 are present in the NSL KDD dataset. The performance of the classifier is checked against them. It consists of 41 attributes, 5 normal classes, and 4 attack types namely Dos, Probe, R2L, and U2R. The experimental result concludes the proposed EGWCSO with Naïve Bayes classifier gives higher precision, recall and accuracy values and less execution time compared to previous algorithms.