The determination of the Convective Heat Coefficients (CHC) is a challenge to get a detailed knowledge of the microclimate of greenhouses in order for instance to improve energy management of these buildings. Computational Fluid Dynamics tools (CFD) have been developed to predict the distributed inside climate. However, most of the previous studies focus on steady state cases for which the boundary conditions are established a priori by the modeller. This point is one of the limitations of CFD use. Complementary, Energy Simulation (ES) is based on an energy balance method which takes account of the dynamic variation of the boundary conditions (i.e. meteorological conditions) all day long, meaning that the wall temperatures are calculated at each time step. In the present paper, an ES model predicting the thermal behaviour inside a closed tunnel greenhouse is coupled to a CFD model. The ES model transmits the wall temperatures to the CFD model and conversely, the CFD model provides the ES model with CHC. The comparison between simulation results shows differences between the CHC calculated by ES (from empirical correlations) and those calculated by CFD. These differences are clearly higher for the vertical walls than for the roofs.