This research explores the applicability of combining clustering technique with stacked generalization for Taiwan housing prices prediction. Taking advantage of the most currently available Taoyuan City Actual Price Registration Data, we first expanded the clustering-based ensemble learning method by Trivedi et al. (2015) to develop two-layer clustering-based stacked generalizers. In the first layer, three machine learning methods (Lasso, KNN and Decision Tree) were used to construct the cluster models. In the second layer, Linear Regression, Random Forest and XGBoost were used to build meta models. These developed stacked generalizers are then used to predict housing prices in the Taoyuan City. Their prediction accuracies are then compared with that from other machine learning methods, including Linear Regression, Random Forest and XGBoost.