Intrusion Detection Systems play a critical role in modern cybersecurity. Cybersecurity is a global concern, and intrusion detection is a key preventive measure for it. Intrusion Detection Systems (IDS) is a preventive cybersecurity measures to monitor and respond to potential cyberattacks. By using machine learning techniques, IDS can be more accurate and detect threats faster in real-time. This paper explores several ML methods (e.g. Decision Trees, Random Forests, Support Vector Machines, Extreme Gradient Boosting) to evaluate their effectiveness in optimizing IDS. This study proposes an optimized IDS framework using XGBoost for robust threat detection and prevention. Comparative analysis of this model has validated its effectiveness over traditional classifiers.