Traditional machine learning -based Intrusion detection systems often consider using single algorithm for IDS classification, which leads to handling bigdata, higher dimensional data difficult. And also the accuracy and other evaluation metrics achieved will not be that good to use those models in modern real -time environments, where even the smallest inaccuracy in detecting an intrusion will cost plenty. In this paper, the proposal of the use of stacking ensemble method as a classifier in IDS is discussed. Proposed framework is evaluated on the CICIDS2017 dataset. Related work on IDS using ML techniques, data pre -processing, the algorithms used in the classification module and the experimental results are presented in this paper. The experimental results achieved are high compared to other previously done works.