The behavior of three distributed parameter mathematical models was investigated in this study with three different surface boundary conditions: (1) equilibrium moisture is instantly reached at the surface, (2) surface moisture content behaves as a first-order process, and (3) diffusive and convective fluxes are equal at the surface of the grain. The three models were numerically solved and fitted to experimental data of soybean grain hydration by the least squares method. The best model was chosen to analyze the behavior of the effective diffusion coefficient inside the grain as a function of hydration time, temperature, and moisture content. The model with the boundary condition of equality of diffusive and convective fluxes at the surface showed the best fitting of experimental data. The effective diffusion coefficient showed considerable variations with moisture content, temperature, and hydration time, and it was well correlated by the equation: Deff=a0+a1X+a2T+a3X2+a4XT+a5T2+a6X2T+a7XT2+a8T3.