The removal of trace carbon dioxide from air was carried out in a two-bed pressure swing adsorption (PSA) packed with 1/16" MS13X zeolite as adsorbent in the depth of L = 0.79 in to rind systematically the effect of design and operating variables on the product concentration C-a1 and the mass exchange efficiency eta(a) defined by eta(a) = (C-a0 - C-a1)/(C-a0 - Ca1Pd/P-a) for the feed concentration. The operation with a shorter cycle time resulted in a higher performance which approached a limiting value at a half cycle time of shorter than 20 min. The individual superficial velocities, U-a and U-d in the adsorption and desorption steps, and the velocity ratio U-d/U-a had a great effect on the product concentration C-a1. On the other hand, pressure in the desorption Step P-d had little effect on the removal performance expressed in terms of the mass exchange efficiency eta(a) while higher pressure in the adsorption step P-d decreased the removal efficiency. This effect of pressure was explained by the decrease in the adsorption coefficient m and intraparticle diffusivity with an increasing pressure P-a. The experimental result agreed well with a simplified model called short cycle time approximation previously proposed by authors when it was applied to a linear isotherm of adsorption and a half cycle time shorter than 20 min. The model provides a distinctive relation between the product gas concentration C-a1 and operating and design parameters by the following equation(1 - eta(a)U(a)/U-d)/(1 - eta(a)) = exp[K(FO)a (L/U-a - L/U-d)]in which K(FO)a is the unified volumetric mass transfer coefficient based on gas phase driving force and inversely proportional to the sum of column pressure (P-a + P-d).