Progress in Machines and Systems Soundaryadevi M, Jayashree L.S 5 1 2016 Analysis of time series data and accurate prediction of future values are among the most challenging tasks that the data analysts face in many fields. Forecasting of energy demands is very essential because both insufficient and excess energy production may lead to a significant reduction of benefits and high storage costs respectively. In order to discover the regularities in dynamic and non stationary data, improved time series forecasting requires a model that combines multiple prediction models. The Ensemble approach performs better than single learning model and discovers the dynamic patterns in Energy time series data. In this paper, we compare the performance of two different Ensemble learning techniques; Bagging (Bootstrap Aggregating) and stacking in forecasting energy time series data. Stacking technique used in this paper, combines different classifiers like Radial Basis Function (RBF), Multilayer perceptron (MLP) and Support Vector Machine (SVM)