In the optimization of radioactive waste disposal facility design, it is important to compare multiple facility design options with the aim of reducing the effective dose as low as reasonably achievable. In this study, a methodology for optimizing the facility design using a probabilistic approach was proposed, and two case studies were provided the application of the methodology and interpretation of the dose distributions. This methodology incorporates the time integration of the 95th percentile value and the sum of coefficients of variation extracted from the probability distribution of dose profile over time as the indicators for relative comparison on the optimization. This methodology enables consideration of the dose with uncertainty for the entire long-term assessment period in the optimization. This allows a more multifaceted comparison of options and is expected to improve the ability to explain optimization with the consideration of long-term uncertainty. While various factors including economic, social, and possibly others are relevant to the decision process, this study focuses on the dose estimations as an indicator for optimization. In the case studies, intermediate-depth disposal in Japan, which is one of the disposal systems of low-level radioactive waste, is used as an example and the groundwater release scenario was the focus. The comparison of multiple options in low-diffusivity and low-permeability layers were assumed. These are the important factors in the migration of radioactive nuclides, while other parameters can also be considered in optimization by using the proposed methodology.