Cybersecurity has gained significant prominence in response to the increasing prevalence of the Internet of Things (IoT), the rapid expansion of computer networks, and the multitude of associated applications. Therefore, identifying various cyberattacks or abnormalities within a network and developing a proficient intrusion detection system, which is indispensable in contemporary security, is progressively significant. Machine learning can build intelligent, data-driven intrusion detection systems. This study examines the application of machine learning for detecting cybersecurity attacks. A Machine Learning (ML) based Intrusion Detection System was developed using data preparation, feature engineering, model training, and validation. This study suggests that various metrics and methodologies can be employed to compare machine learning models, depending on the type of classification. Machine learning-based attack detection offers a strong solution against cyber threats. The K-Nearest Neighbour model achieves 92% accuracy, 97% precision, and 87% recall, outperforming other models. These methods utilize algorithms to recognize patterns and anomalies, and extensive datasets train models to identify and respond to complex attacks in real-time. Machine learning enhances cybersecurity by continuously learning from new data, enabling a dynamic defense against evolving threats.