Cyber threats are constantly changing, they put digital infrastructure at risk. This research uses XGBoost, a cutting-edge machine learning algorithm, to improve the capabilities of cyber threat detection. We hope to create a threat detection system that has both high accuracy and effectiveness by utilizing XGBoost's remarkable performance in managing sizable, complex datasets and its capacity to identify patterns that are complex in network transactions. A huge dataset of network traffic, that includes both benign and malicious activity, will be used to train the suggested system. We will evaluate the accuracy of the model in classifying cyber threats, such as malware attacks, phishing attempts, and advanced persistent threats, through thorough evaluation as well as testing. Once this XGBoost-powered system is successfully deployed, businesses will have a strong impact.