Development of effective heat treatments is crucial to achieve food products' safety, and predictive microbiology is an excellent tool to design adequate processing conditions. This work focuses on the application of a modified Gompertz model to describe the inactivation behaviour under time-varying temperature conditions at the surface of a food product. Kinetic studies were carried out assuming two different heating regimes, typically used in surface pasteurisation treatments, and compared with isothermal conditions. Parameters were estimated on the basis of generated pseudo-experimental data. It was concluded that the heating period greatly affects microbial inactivation and parameter estimation. If a slow heating treatment is used, the process time should be extended to achieve a given microbial load when compared to a fast heating process. This is explained by the fact that, in the slow heating rate process the temperature was below the lowest temperature for inactivation for a much longer time, in comparison with the fast heating regime. (c) 2005 Elsevier Ltd. All rights reserved.