Solar radiation significantly impacts the energy received from the sun in a specific area, crucial for planning non-conventional renewable energy power plants like solar photovoltaic or solar thermal systems. Variability in this resource, influenced by climate and geography, poses challenges for solar integration planning. Numerical models estimate solar resource but lack realtime and future responses. Machine Learning (ML) offers heuristic predictive tools, using extensive datasets and algorithms for quantifying and forecasting solar radiation. A proposed ML model incorporates geolocation and links primary resource with climate data from diverse Colombian cities. It consists of three stages: clustering, estimation, and response, utilizing ML predictors selected by criteria and literature review. Model response is validated using statistical methods, providing accurate solar resource predictions.