Design and optimization of thermal processing of foods need accurate dynamic models to ensure safe and high quality food products. Transfer functions had been demonstrated to be a useful tool to predict thermal histories, especially under variable operating conditions. This work presents the development and experimental validation of a dynamic model (discrete transfer function) for the thermal processing of tuna fish in steam retorts. Transfer function coefficients were obtained numerically, using commercial software of finite elements (COMSOL Multiphysics) to solve the heat transfer balance. Dependence of transfer function coefficients on the characteristic dimensions of cylindrical containers (diameter and height) and on the sampling interval is reported. A simple equation, with two empirical parameters that depends on the container dimensions, represented the behavior of transfer function coefficients with very high accuracy. Experimental runs with different size containers and different external conditions (constant and variable retort temperature) were carried out to validate the developed methodology. Performance of the thermal process simulation was tested for predicting internal product temperature of the cold point and lethality and very satisfactory results were found. The developed methodology can play an important role in reducing the computational effort while guaranteeing accuracy by simplifying the calculus involved in the solution of heat balances with variable external conditions and emerges as a potential approach to the implementation of new food control strategies leading not only to more efficient processes but also to product quality and safety.