Abstract: Disposal of radioactive waste is a major issue related to land and environmental protection, public safety, nuclear industry health, and sustainable development. Safety System Analysis (SSA) is an important approach to ensure the safety of radioactive waste disposal facilities throughout the entire process, from site selection and construction to operation and closure. Uncertainty analysis is a crucial component of SSA. Uncertainties stemming from uncontrollable external factors such as environmental changes and human behavior can have significant impacts on the safety of radioactive waste disposal facilities, necessitating uncertainty assessment. Latin Hypercube Sampling (LHS) is a commonly used method in uncertainty analysis. However, LHS alone cannot directly address the requirement for considering correlations between input parameters. When there are certain correlations among input parameters, other methods are needed for improvement. In this article, an improved approach to LHS using simulated annealing is adopted. Simulated annealing is a heuristic algorithm inspired by the annealing process in metallurgy, which simulates the slow cooling of materials to reach the lowest energy state. By iteratively updating samples, this method achieves more representative and efficient sampling.