The dependence on digital infrastructure is growing and so are cyber security threats, organizations are in need of intrusion detection. Because of their rigidness, conventional definitions of rule-based security systems miss out on cyber threats. Accordingly, new approaches such as those using machine learning have been developed to enhance the security level. So, there is a need for developing an Intrusion Detection System as a machine learning model to monitor network traffic to identify the suspicious activities. Various machine learning models are trained and tested on most popular dataset used to intrusion detection that is KDD Cup 1999 dataset. These key steps include data processing and feature selection, and model training, as well as performance evaluation. To select the optimal model, different classification algorithms are considered such as Decision Tree, Random Forest, k-Nearest Neighbor, and Neural Networks in this study. Random Forest outperformed in terms of accuracy implementing Explainable Artificial Intelligence methods to provided explanation to the decision making process of the model. Results indicate that Machine Learning based Intrusion Detection System can greatly improve the accuracy of detection with a major reduction in the false positives, hence their improving the defenses in cybersecurity and efficiency in mitigation of threats in real time.