Sustainability has emerged as a critical focus in today’s world, with solar energy emerging as a viable power source, particularly for developing countries like Bangladesh that face substantial energy demands. While global initiatives are underway to harness solar power more efficiently through artificial intelligence, its application remains limited in Bangladesh. This chapter employs random forest (RF) and feed-forward neural network (FNN) models to forecast continuous hourly solar radiation, utilizing raw data collected from a station in Feni, Bangladesh. Through comprehensive time-series data processing, feature selection, and hyperparameter optimization, we achieved an RMSE of 0.0459 for RF and 0.0395 for FNN. Additionally, our methodology was applied to forecast hourly irradiance in three other cities in Bangladesh, yielding satisfactory results. By incorporating meteorological parameters, time features, and historical time-series data, we established a robust forecasting model that can be applied across Bangladesh, utilizing NASA Power Project’s public database for continuous real-time hourly irradiance forecasts.