The reported methods to extract parameters for solute transport analysis are hard to handle with inconsistent test procedures. A new model with regard to the nonlinear adsorption behavior was defined on the basis of mass percent of solute on the soil particles. The relevant governing equations were deducted and a code to predict the solute transport parameters using backward analysis method was developed under a least square principle. The new method was verified by testing the diffusion behavior of Pb(II) in Kaolin column. The results showed that the nonlinear diffusion theory could better explain the experimental phenomena than the linear diffusion theory. The adsorption of Pb(II) on Kaolin appeared fitting Freundlich model with predicted adsorption capacity at 0.00177~0.00236. The diffusion coefficient of Pb(II) in Kaolin which was determined to be in the range of 0.0868~0.1058 cm2/d decreased with increasing solute concentration. Kaolin was found applicable in barring the diffusion of Pb(II) in spite that the adsorption capacity toward pollutant would be decreased with increasing dry density.