With the rise of demand for synthetic diamond (zirconia) in the gemstone industry, it is important to predict price for zirconia as it is vital for manufacturers, suppliers, and buyers. In the current scenario, machine learning plays an important role to access the feature of stones and predict the price. This study does an exhaustive evaluation with various machine learning regression models including both individual models and blended models. Total nineteen models are being used for performance comparison including standalone models and blended models. The data is being pre-processed using categorical encoding, outlier removal, and standardization for minimizing underfitting and overfitting. The study suggests LightGBM, random forest, and Extra Trees give overall best performances compared to other models. To improve prediction performance, we ensemble the top-performing models to further improve the generalization and accuracy. This research highlights the strength of blended models that can offer a scalable tool for price prediction of a gemstone.