Underground shaft excavation in the Himalayan region is a very challenging task for rock engineers due to its complex geological conditions. The rock mass heterogeneity that prevails in the Himalayan region increases uncertainty in the identification of rock mass quality class. Thus, the use of empirical approaches alone may not be able to make good prediction capabilities. In this context, the development of a probability approach is of utmost importance for uncertainty assessment of rock mass quality class in shaft excavation. In this manuscript, uncertainty assessment is carried out using Monte Carlo simulation analysis for the prediction of rock mass quality class where two ensemble machine learning techniques such as Bagging and XGBoost have been introduced. The results show that mapped input parameters for Q-system of rock mass classification have considerable variations. In the ensemble machine learning technique, XGBoost model shows good prediction performance as compared to the Bagging model. Thus, Monte Carlo simulation-based ensemble machine learning technique can be used for rock mass quality class prediction during planning and design stage of shaft excavation in the Himalayan region since the prediction approach may help to minimize the risk of geological uncertainty and enhance shaft excavation performance.