This paper deals with the determination of expressions to estimate the thermal diffusivity of foodstuffs placed in cans during their pasteurisation. To this end, experimental data of the temperature in the geometric centre of the can were compared with data obtained by simulations of heat penetration. For these simulations, the diffusion equation was numerically solved by using the finite volume method, with fully implicit formulation. The numerical solution was coupled to an optimiser based on the inverse method, in order to fit this solution to the experimental data obtained during the pasteurisation of tomato puree and custard cream placed within sealed cans. In the literature, the assumption of constant thermal diffusivity during the pasteurisation is common but for the analysed cases such hypothesis resulted in biased fits. Better results were obtained supposing an increasing expression for the thermal diffusivity, as function of the local temperature. In this case, the errors of the experimental data about of the numerical solution can be considered randomly distributed.