Several methods have been proposed currently for evaluating the crack width of a mortar specimen. Among these, the water permeability test is widely used to estimate crack width because water permeability is directly related to the average crack width of a specimen through which water passes. However, the viscosity of water makes precise crack width measurement challenging. The possible inflow (outflow) of foreign (healing) substances could affect the test results. To circumvent this limitation, this study proposes a gas diffusion test using oxygen rather than water as the medium. The proposed method includes a process that could compensate for gas diffusion from specimen parts other than the crack, allowing for a more precise estimation of crack width. The crack width can indeed be estimated with an error of 4% or less.