The accurate simulation of stomatal behavior in diverse soil moisture conditions is important for the characterization of the responses and the adaptive mechanisms of vegetation ecosystems to climate change and for the prediction of the carbon and water cycles between the plants and the atmosphere in the context of climate change. Based on the leaf gas exchange parameters data for spring maize from a field progressive drought manipulation experiment, the applicability of four common stomatal conductance models (Jarvis, BWB, BBL, and USO) in spring maize were studied under drought conditions, the effects of the soil water response function on the stomatal conductance models were evaluated, and a suitable soil moisture range for each stomatal conductance model was discussed. The results revealed that the simulation accuracy of the Jarvis, BWB, BBL, and USO models was affected by the soil relative water content (SRWC). In progressive drought conditions, the BBL model performed the best, followed by the USO model and the BWB model, and the Jarvis model performed the worst. After the introduction of the soil water response function, the modified USO model performed the best, followed by the modified BBL and BWB models, and the modified Jarvis model performed the worst. The simulation accuracy of the BWB and USO models was improved by the introduction of the soil water response function; the normalized root mean square error (NRMSE) value decreased by 1.99% and 1.02%, respectively, and the relative error (RE) value decreased by 3.20% and 0.63%, respectively. Although the performance of the Jarvis and BBL models decreased after the introduction of the soil water response function, the NRMSE value increased by 4.70% and 3.45%, respectively, and the RE value increased by 6.02% and 2.00%, respectively. The residual stomatal conductance of the BBL and USO models was relatively smaller than that of the Jarvis and BWB models, which indicated that the BBL and USO models displayed considerable stability under progressive drought conditions. According to the relationship between stomatal conductance and SRWC and in terms of the 95% confidence intervals, the Jarvis, BBL, and USO models were applicable when SRWC was between 33% and 83%, whereas the BWB model was applicable when SRWC was between 33% and 76%. After the introduction of the soil water response function, the modified BWB model could be applicable for the current experimental soil moisture range. The results might provide references for the selection of suitable stomatal conductance models to improve the efficient use and assessment of agricultural water resources.