Abstract: The capillary rise is capable of revealing the mechanism involved in groundwater evaporation.Among the parameters of this experiment,the maximum rate of groundwater evaporation( E_max) is an indispensable parameter in the calculation of groundwater evaporation,which can be estimated by the Gardner model. When using the Gardner model,it has widely been accepted to ignore parameter b during hydraulic conductivity prediction for simplification purpose,which may yet affect the prediction accuracy.Therefore,an investigation on how b affects E_maxis necessary. Loam soil samples from Yangling,Changwu,and Linyi regions were selected for indoor capillary rise and one-dimensional evaporation experiments,with two solutes（ KCl and NaCl） that were dissolved in groundwater at three concentrations（ 1,5,and 30 g/L） and one control treatment（ CK） of distilled water without any salt solutions. Then E_max was calculated by the observation of the capillary rise and the sensitivity of b in the Gardner model was evaluated. Results indicated that the capillary rise rate of the treatments supplied by salty groundwater was larger than that supplied by distilled water,which further led to a larger E_max; besides,the capillary rise rate and E_maxdecreased with the increase of the salt concentration when the concentration of KCl and NaCl solutes was in the range of 1-30 g/L. Furthermore,during the prediction of the Gardner-model-based E_max,ignoring b and considering b always caused an overestimation and underestimation,respectively.And the Gardner model considering b had higher accuracy for E_maxprediction,with the relative error of1. 16%-4. 50%. Moreover,the influence of ignoring b on E_maxgradually weakened with the increase of the b value. Improving the accuracy of E_maxprediction can benefit the decision-making of cropland irrigation schemes and the understanding of groundwater evaporation mechanism.