Network Intrusion Detection Systems (NIDS) are essential for enhancing network security. Traditional research has often relied on the KDDCUP99 dataset; however, the UNSW-NB15 dataset provides a more comprehensive framework for evaluating NIDS. This study investigates the performance of seven machine learning models: Support Vector Machine (SVM), Random Forest, Logistic Regression, Linear Regression, Multi-layer Perceptron (MLP) Classifier, K-Nearest Neighbors, and Decision Tree for both binary and multiple class classification. Results indicate that Random Forest tops the accuracy with 98.63% in binary classification, while the MLP classifier excels in multi-class classification with 89.93%.