The spatial distribution of constituents in concrete affects the material performance, and voids present in cast concrete significantly affect permeability and durability because most of the air and water are passed through voids inside the material. Therefore, a proper investigation of the void distribution inside concrete is important to evaluate and understand material characteristics. To visualize and investigate the spatial distribution of voids in a concrete specimen, X-ray CT images and probabilistic description methods including reconstruction were used in this study. Additionally, a virtual concrete specimen was reconstructed using the probabilistic information of real specimens, and the permeability of each specimen was evaluated using finite element simulation, and compared with experimental data. From the results, we confirm that the proposed reconstruction method can effectively generate a virtual cement paste specimen with a specific void distribution.