An improvement of the Piringer model, allowing the prediction of a worst case migration from packaging to food is presented here. The authors are proposing other constants for the calculation of the upperbound value of the diffusion coefficient, using experimental data determined by a film to film method. Considering the plasticizing effects of food simulants, a model involving the variation of the diffusion coefficient versus space and time must be used. Future fields of investigation are discussed: the relationship between diffusion coefficients and the volume of the migrant (instead of molar mass), and the variation of diffusion coefficient activation energy with temperature.