The responses of CO2 assimilation to [CO2] (A/Ci) were investigated at two developmental stages (R5 and R6) and in several soybean cultivars grown under two levels of [CO2], the ambient level of 370 μbar versus the elevated level of 550 μbar. The A/Ci data were analyzed and compared using various curve-fitting methods by either the combined iterations or the separated iterations of the Rubisco-limited photosynthesis (Ac) and/or the RuBP-limited photosynthesis (Aj) using various curve-fitting methods: the linear 2-segment model; the non-rectangular hyperbola model; the rectangular hyperbola model; the constant J method and the variable J method. Inconsistency was found among the various methods for the estimation of the maximum rate of carboxylation (Vcmax), the mitochondrial respiration rate in the light (Rd) and mesophyll conductance (gm). The analysis showed that the inconsistency was due to inconsistent estimates of gm values that decreased with an instantaneous increase in [CO2], and varied with the transition Ci cut-off between Rubisco-limited photosynthesis and RuBP-regeneration-limited photosynthesis, and due to over-parameters for non-linear curve-fitting with gm included. We proposed an alternate solution to A/Ci curve-fitting for estimates of Vcmax, Rd, Jmax and gm with the various A/Ci curve-fitting methods. The study indicated that down-regulation of photosynthetic capacity by elevated [CO2] and leaf aging was due to partially the decrease in the maximum rates of carboxylation and partially the decrease in gm. Mesophyll conductance lowered photosynthetic capacity by 18% on average for the case of soybean plants.