The design of adsorption-based separation processes using novel adsorbents requires reliable data for the adsorption of fluid mixtures on candidate adsorbents. Due to the difficulty of generating sufficient data across possible operating conditions, process designs generally rely on interpolation of pure-component data using a model, most commonly ideal adsorbed solution theory (IAST), and related theories. There are many cases where IAST fails to provide an adequate description of mixture adsorption, usually due to the fact that practical adsorbents do not have uniform surfaces. We have evaluated the use of a segregated version of IAST, where competition is assumed to occur at isolated adsorption sites. This simple modification can provide the correct description of adsorption across a large range of pressures using ideal isotherm models. We also demonstrate the importance of identifying multiple sites even for weakly adsorbing components to provide the correct behavior at high pressure. (c) 2013 American Institute of Chemical Engineers AIChE J, 59: 3054-3064, 2013