Diffusion coefficient is usually used to evaluate the methane (CH4) diffusion properties in the coal matrix and is vital to coalbed methane (CBM) development. Although extensive literature on the CH4 diffusion coefficient can be obtained, most of them aim at the whole coal or coal rank instead of the macrolithotype. Additionally, the primary structure of coal was destroyed with the common determination technologies (e.g., the particle, steady-state, and inverse diffusion methods) which could result in great errors. In this work, to avoid the shortcomings of the above methods, nine flake coal samples from six coal mines in the Qnshui and Ordos Basins were prepared to determine the CH4 diffusion coefficient with the slab calculation model. Meanwhile, the effects on the diffusion from the gas pressure, temperature, water saturability, and coal pore structure, and the gas adsorption capacity controlled by the coal rank and macrolithotype, were analyzed to reveal the diffusion mechanism (mode) at the CBM reservoir and laboratory conditions. Results show that the CH4 diffusion coefficient, at an order of magnitude of 10(-10) u m(2)/s measured with the flake coal sample, is more truthful. High temperature and gas pressure, low water saturability, developed pore structure, and high gas adsorption capacity contribute to large CH4 diffusion coefficient. Although the higher rank coal has the larger gas adsorption capacity, the CH4 diffusion coefficient exhibits a "U" shape (first decreasing and then increasing) with the increase of coal rank due to more micropores in low- and high-rank coals than the middle-rank coal. From the bright to dull coals at the same coal rank, the decreasing development of pore structure and gas adsorption capacity causes the decreasing CH4 diffusion coefficient. But compared to the coal rank, the influence of coal macrolithotype on CH4 diffusion coefficient is weaker. In addition, the CH4 diffusion modes in coal mainly are transitional and Fick diffusions in the CBM reservoir and laboratory.