In our previous study, we developed a CFD thermal analysis model for a CANDU spent fuel dry storage silo. The purpose of this model is to reasonably predict the thermal behavior within the silo, particularly Peak Cladding Temperature (PCT), from a safety perspective. The model was developed via two steps, considering optimal thermal analysis and computational efficiency. In the first step, we simplified the complex geometry of the storage basket, which stored 2,220 fuel rods, by replacing it with an equivalent heat conductor with effective thermal conductivity. Detailed CFD analysis results were utilized during this step. In the second step, we derived a thermal analysis model that realistically considered the design and heat transfer mechanisms within the silo. We developed an uncertainty quantification method rooted in the widely adopted Best Estimate Plus Uncertainty (BEPU) method in the nuclear industry. The primary objective of this method is to derive the 95/95 tolerance limits of uncertainty for critical analysis outcomes. We initiated by assessing the uncertainty associated with the CFD input mesh and the physical model applied in thermal analysis. And then, we identified key parameters related to the heat transfer mechanism in the silo, such as thermal conductivity, surface emissivity, viscosity, etc., and determined their mean values and Probability Density Functions (PDFs). Using these derived parameters, we generated CFD inputs for uncertainty quantification, following the principles of the 3rd order Wilks’ formula. By calculating inputs, A database could be constructed based on the results. And this comprehensive database allowed us not only to quantify uncertainty, but also to evaluate the most conservative estimates and assess the influence of parameters. Through the aforementioned method, we quantified the uncertainty and evaluated the most conservative estimates for both PCT and MCT. Additionally, we conducted a quantitative evaluation of parameter influences on both. The entire process from input generation to data analysis took a relatively short period of time, approximately 5 days, which shows that the developed method is efficient. In conclusion, our developed method is effective and efficient tool for quantifying uncertainty and gaining insights into the behavior of silo temperatures under various conditions.