Abstract- Determining the correct secondary structures of protein is one of the key roles in predicting the right folded shape, which is used for determining gene function. In this study, we proposed a technique to predict secondary protein structures based on ensemble learning. This study uses protein residue from the enzyme data repository. Position-specific scoring matrix (PSSM) profile Combined with the physicochemical feature are used several weak classifiers with lower accuracy yet lower complexity including Naïve Bayes, k-Nearest Neighbor, and decision tree is a technique that combines several ensemble learning techniques to achieve better prediction performance whit combined bagging, boosting, and stacking steps to improve the performance. The results show that our proposed method outperforms individual classifier and individual ensemble technique.