This study investigates enhancing energy security in isolated areas of Mindanao through advanced machine learning techniques, specifically the NBEATSx model, to forecast solar irradiance and wind speed. It aims to evaluate renewable energy potential accurately by utilizing localized meteorological data and integrating exogenous variables. The primary objectives include demonstrating the model’s predictive accuracy, estimating power generation potential and the significance of geographical characteristics in renewable energy planning. The NBEATSx model achieved high predictive accuracy, with R2 values exceeding 0.80 for solar and wind power predictions. Estimated annual power generation ranges from 103,657.07 kWh to 928,412.19 kWh for wind energy and from 9,326.83 kWh to 14,490.14 kWh for solar energy. These results underscore the feasibility of renewable energy as a power source for isolated regions and the importance of site-specific assessments. This study contributes to renewable energy resource assessment by demonstrating advanced machine learning models for precise predictions.