Controlling the quality of industrial products requires an accurate comprehension of the material behavior during the several transformation phases. The exact modeling of the heat transfers taking place is crucial for manipulating rigorously the manufacturing processes. Though, this modeling necessitates the knowledge and the conform characterization of the material thermophysical properties which are the thermal conductivity, the specific heat and the specific volume. Currently, these properties are well defined in the solid state, however they are less mastered in the liquid state (for thermoplastics) and during transformation. The aim of this paper is to establish an inverse heat transfer method for predicting the variation of the thermophysical properties during the polymer transformation. This phase change corresponds mainly to the crystallization of the molten polymer to forma solid with an ordered internal arrangement of molecular chains. The concept of the inverse method consists in minimizing the difference between the numerical outputs obtained by a developed finite difference model and the temperature evolution of the material during its transformation. A hybrid optimization algorithm combining a stochastic algorithm with a deterministic one is adopted and its robustness is verified by using synthetic noisy data generated by the numerical model. A sensitivity analysis is conducted in order to test the feasibility of the parameters simultaneous identification. The expected results will allow to describe the variation of the thermophysical properties function of temperature T and the relative crystallinity α.