Molecular diffusion is the dominant transport mechanism in the release of radionuclides from engineered barriers of a low-level radioactive waste repository. To evaluate the release rate, proper measurement of the diffusion coefficient for each radionuclide is essential. Since several diffusion coefficients have been used in the literature, they are clearly defined in this paper to avoid ambiguity. Theoretically, to obtain a good estimate of the intrinsic diffusion coefficient, it is required that the diffusion process approach a steady state, which cannot be confirmed readily. In fact, both the apparent diffusion coefficient and the capacity factor of the porous medium vary during the transient period. Therefore, we propose a numerical technique to treat the transient period and the steady-state period separately. By analyzing the experimental data measured at all stages of the diffusion process, this method can determine the apparent diffusion coefficient and capacity factor in the transient period as well as the intrinsic diffusion coefficient in the steady-state period. The results of this method are compared with those of the conventional graphical method for the diffusion of L3q-and tritiated water (THO) through cement. The data-plotting work makes the graphical method cumbersome and may be inaccurate in some cases. The proposed method can readily provide reliable analyses of experimental data with measured errors. Furthermore, this study is able to interpret the hypothesis of "dead-end" pores in the transient period, and the influence of the effective porosity in porous media with different thicknesses. Our method can also judge whether the diffusion experiments have been executed long enough to reach steady states.