Two approaches exist for conjucting solar radiation: Through logical estimations and machine learning standards. Main objective of this endeavor inhabit to display a outline of solar energy expectation utilizing machine learning calculations in this context. By using NASA’s geo-satellite database and NREL’s Solar Radiation Research Laboratory. The economical utilize of unreservedly accessible sun powered radiation as a renewable vitality source depends on exact predictive models to quantitatively survey future vitality potential .We are utilizing different machine learning (ML) calculations, such as lightgbm, gradient boosting, XGBoost, linear regression, and LSTM in expansion to that we are utilizing ensemble algorithms have been utilized for the predictions. Compared to other calculations, the “Gradient Boosting” with 0.99 of R \(^2\) and ensemble of “Gradient Boosting and lightgbm” performed better. The hybrid model implements a sophisticated interaction protocol with LSTM Network which captures complex temporal dependencies in solar radiation data and LightGBM which performs feature importance ranking and non-linear transformations