Renewable energy is a green and low-carbon energy source, which is of great significance for improving energy structure, protecting the ecological environment, addressing climate change, and achieving sustainable development. This article investigates the application of ensemble learning models in the prediction of renewable energy increment. Traditional prediction methods have problems due to their own limitations, such as imbalanced data and rough feature selection, resulting in low model prediction accuracy. This article proposes an ensemble learning based energy increment prediction model (ERLSK model). The model includes four traditional machine learning classification models: KNN, random forest, SVM, and logistic regression. The best hyperparameter of each model are obtained by cross validation and grid search. By using different feature selection and model training methods, different prediction results are obtained, and these results are weighted and fused to obtain the final classification result. The experimental results indicate that the model is effective and feasible.