With the rapid expansion of high-speed internet, the demand for secure and resilient networks has significantly increased. In recent years, network security has become a critical component of organizational information systems, driven by the growing complexity and frequency of cyberattacks. Network Intrusion Detection Systems (NIDS) have emerged as a fundamental line of defense to ensure data integrity and system availability. The advancement of machine learning techniques has notably enhanced the performance and accuracy of NIDS, enabling faster and more precise detection of threats. This study presents a machine learning-based approach for identifying cybersecurity threats using real network traffic data from the CICIDS2017 dataset. The proposed model employs the eXtreme Gradient Boosting (XGBoost) classifier to effectively differentiate between benign and malicious network flows. Experimental results demonstrate a high prediction accuracy of 99%, confirming the model’s effectiveness in real-world scenarios. The integration of intelligent algorithms with real-time threat intelligence proves to be a promising direction for developing robust and adaptive cybersecurity defense mechanisms. This research contributes to the ongoing efforts to leverage artificial intelligence in enhancing network security and provides a foundation for future improvements in automated threat detection systems.