Zero-dimensional zonal models are seen as interesting tools for engine simulation due to their simplicity and yet accuracy in fitting or predicting experimental data. For combustion, a common model is a dual zone model, in which two-zones, spatially homogeneous, are set during the combustion process. Such model take into account an interface of infinitesimal thickness for the separation between zones. The success of this simulation approach depends on the accuracy of the heat transfer model. These models aim to obtain the heat transfer coefficient from the combustion gases in contact with the cylinder walls. Several heat transfer correlations from the literature can be used to obtain the heat transfer coefficient. To evaluate which of these correlations is able to better fit the experimental data of in-cylinder pressure, a comparative study between Eichelberg correlation, who assumed natural convection of combustion gases, and correlations of Woschni, Hohenberg, Sitkei and Annand, which adopted the forced convection, were performed. Comparisons were made using the software Matlab® for an internal combustion engine using different fuel blends: hydrous ethanol fuel (HEF) and mixtures of ethanol and water in the proportions of 90% (E90W10), 80% (E80W20), 70% (E70W30) and 60% (E60W40) by volume of ethanol. This Part B paper shows tests results with maximum brake torque timing for wet ethanol. The results have shown that correlations of Hohenberg and Annand are most suitable for the type of engine and fuels considered in this study, both being able to predict accurately the pressure curve inside the cylinder.