Catalytic performance is influenced by the effective gas diffusivity in a heterogeneous catalyst. The diffusivity is dependent on the microstructure of the catalyst. Therefore, the microstructure of a catalyst must be characterized, and the effective gas diffusivity should be predicted using the microstructural properties. Mercury intrusion porosimetry (MIP) is a popular technique to characterize microstructure and subsequently predict the effective gas diffusivity with the random pore model (RPM) using the MIP results. However, this methodology cannot be used for multilayered catalysts such as automotive catalysts due to the difficulty of microstructure characterization. Three-dimensional (3D) image analysis is expected to be an informative technique to characterize the porous structure of multilayered catalysts because each layer can be analyzed independently. In this research, Xray nano computed tomography (nano-CT), 3D-transmission electron microscopy (TEM) and MIP were employed to characterize porous Al2O3 pellet samples. A series of samples with different microstructures were prepared. The effective diffusion coefficient of CO2 was evaluated experimentally for each sample. The effective diffusion coefficient was also predicted with the RPM or by simulation. Pore properties obtained by MIP were applied for the RPM. 3D images obtained by nano-CT and 3D-TEM were used for application of the simulation. A strong correlation between the experimentally evaluated effective diffusion coefficient and the predicted diffusion coefficients was thus confirmed.