Nitrogen dioxide (NO2) is a significant air pollutant primarily emitted from traffic and industrial activities, posing health risks. Accurate predictions of urban NO2 concentrations are essential for effectively controlling air pollution. In this study, we focus on forecasting NO2 levels in New Borg El-Arab City, Alexandria, Egypt — a rapidly developing industrial area — to enhance air quality management and urban planning. This research employs comparative analysis of three machine learning (ML) models, including Artificial Neural Networks (ANN), Random Forest (RF), and Support Vector Machines (SVM). Hourly datasets were collected from the New Borg El-Arab City Weather Station and an IoT-based air quality monitoring system with Arduino from 2nd January 2021 to 30th May 2021. While Key environmental and meteorological variables, such as Sulfur Dioxide (SO2), Fine Particulate Matter (PM2.5), Temperature (T), Relative Humidity (RH), and Wind Direction (WD), were collected, only four variables were selected to forecast NO2 concentration based on their higher correlation with NO2 as determined using Correlation Matrix. The study employed R2, RMSE, MAE, and MSE as evaluation metrics to assess the model's performance, ensuring robust comparisons. The findings indicate that ANN, RF, and SVM achieved a high accuracy, exceeding 91% for NO2 prediction. The comparative analysis revealed that the ANN surpassed the other ML models with an RMSE of .7350 during training and 1.2281 for testing. This study contributes to the ongoing efforts to achieve sustainable urban development and improve public health outcomes in Egypt.