A comparative analysis of the results Solar system’s output power predictions is required for the power grid to work properly and for the solar system’s ability to regulate its energy flows as efficiently as possible. Before projecting the solar system’s output, the prediction must be centered on solar irradiance. The two most effective strategies for global solar radiation prediction are machine learning and physical models. Many studies have been undertaken on solar radiation forecasting. The disadvantages of current techniques are that they are ineffectual, imprecise, and time demanding. In this study, we forecast sun radiation using Random Forest Regression (RFR), Gradient-Boosted Decision Tree (GBDT), Support Vector Machine (SVM), and Artificial Neural Network Regression. A comparison of results to determine the number of ensemble approaches can be used to increase prediction accuracy in solar radiation recasting. According to our experimental results, our ensemble approach achieves 93.1% accuracy of predictions on the training dataset and 91.4% on the testing dataset. Also, for 10,000 nodes, the forecast time is less than 50.6 ms.