This research examines three machine learning models the LASSO-RFR Hybrid, LASSO standalone, and Random Forest Regressor standalone models—for forecasting solar energy generation in Abuja, Nigeria. The hybrid model utilizes the advantages of both LASSO and Random Forest to enhance prediction accuracy across varying weather patterns. Although the LASSO-only model showed superiority in certain instances, the hybrid model typically delivered consistent predictions favourably predicting outcomes when essential weather variables such as temperature and cloud cover were excluded. Based on the findings, the hybrid model stands out as an option for forecasting solar energy in Abuja in scenarios where only solar, temperature and cloud cover variables are available for analysis. It also shows competitive performance in its adeptness in managing varying weather conditions and delivering precise predictions.