This paper presents a global sensitivity analysis based strategy for uncertainty quantification of the operating revenue and the carbon dioxide emissions in a crude oil distillation column. Aspen HYSYS is used for rigorous simulation of the column. A two-stage approach, which is executed by an interface between MATLAB and HYSYS, is implemented in this study. A multiplicative dimensional reduction method is applied to first identify which factors have the most influential to the model outputs, e.g. the operating revenue and the CO2 emissions. In the second stage, the uncertainty quantification regressed by Gaussian prediction regression method is exploited. As a result, there is a good agreement between the predicted results by the Gaussian prediction regression method with that of the conventional Quasi-Monte Carlo approach, which shows that the computational efforts was reduced significantly compared to the conventional study (about less than 10 times). Furthermore, the Kernel points estimated by the Gaussian prediction regression method are generated to highlight influential factors identified in the first stage.