To accurately describe the adsorption equilibria of N 2 and O 2 on LiLSX zeolite in the design of air separation by pressure swing adsorption, an applicable model (aNRTL−RAST) was developed. It was based on the real adsorbed solution theory (RAST) in which the activity coefficient was calculated by the nonrandom two-liquid theory (aNRTL). The binary adsorbate− adsorbate interaction parameters of the aNRTL theory were estimated from the pure component isotherms. The aNRTL−RAST model was used to predict the mixed gas equilibrium data under different adsorption conditions and compared with the extended Langmuir (E-L), two-dimensional equation of state (2D-EOS), and ideal adsorbed solution (IAST) models. aNRTL−RAST showed better accuracy than the other models, and the average errors for different adsorption conditions were <5%. Therefore, this model can be used to predict the adsorption equilibrium of mixed gases, avoiding complicated and tedious experiments. This study provides a theoretical basis for the selection of oxygen adsorbents for the separation process design of industrial air separation.