Land Surface Models (LSM) have shown some difficulties to properly simulate day-time 2-m air and surface skin temperatures. This kind of models are coupled to atmospheric models in mesoscale modelling, such as the Regional Atmospheric Modeling System (RAMS) and the Weather Research and Forecasting (WRF) Model. This model coupling is used within Numerical Weather Prediction Systems (NWP) in order to forecast key physical processes for agricultural meteorology and forestry as well as in ecological modelling. The current study first evaluates the surface energy fluxes and temperatures simulated by these two state-of-the-art NWP models over two distinct vegetated covers, one corresponding to a poor and sparsely vegetated area and the other one corresponding to the tall and well-vegetated area of a forest. On the other hand, the importance of parameterizing the thermal roughness length within the LSM coupled to the corresponding atmospheric model is also evaluated. The LEAF-3 LSM is used within the RAMS modelling environment while the Noah-MP LSM is applied within WRF. Results indicate that the original version of the models underestimates the temperature during the day, more remarkably in the forested area, whereas modifications in the thermal roughness length successfully simulates the temperature and sensible heat flux forecasts over this area. This study highlights the key role of the surface exchange processes when coupling land and atmosphere models. In this regard, incorporating an extra resistance in the surface-layer parameterization through the thermal roughness length is essential to simulate well both temperatures and sensible heat fluxes, which becomes more relevant over tall and well-vegetated areas, such as a forest. This extra resistance for heat exchange prevents effective molecular diffusion in the layer between the momentum roughness length and the thermal roughness length. Additionally, an appropriate description of the canopy height permits to apply an improved surface-layer formulation over different land and vegetation covers.