The focus of this article is on CO2 capture from natural gas under humid conditions using CALF-20. Configurational-Bias Monte Carlo (CBMC) simulations are used to investigate adsorption of CO2/CH4/H2O mixtures in CALF-20 with varying relative humidities in the bulk gas phase. The CBMC simulations reveal that the mixture adsorption equilibrium shows strong deviations from the estimations of the Ideal Adsorbed Solution Theory (IAST). CBMC simulations for mixture adsorption of the constituent binary pairs, CH4/H2O, and CO2/H2O, reveal that the origin of thermodynamic non-idealities stem from segregated nature of the adsorbed phase and inhomogeneous distribution of guest adsorbates within the pore landscape of CALF-20. The extent and nature of the segregation depends strongly on the % relative humidity of the bulk gas mixture.The important message that emerges from this investigation is the need to incorporate the Real Adsorbed Solution Theory (RAST) for quantitative modelling of fixed-bed adsorbers in natural gas purification with CALF-20.