Urbanization leads to the replacement of natural surfaces with buildings and paved surfaces. This change in surface characteristics together with human activities in urban environments alters heat, moisture and momentum exchange processes in the atmospheric boundary layer and distinguishes the urban climate from that of surrounding rural areas. In the last few years, several efforts have been made in order to improve the representation of urban surface characteristics in mesoscale models, either to the ‘dynamical’ part (impact on the wind field and the turbulent kinetic energy) or to the ‘thermal’ part (impact on the heat fluxes). Shadowing and radiative trapping effects are also considered. In the present study, modifications both to the ‘thermal’ and ‘dynamical’ part were incorporated in the meteorological PSU/NCAR Mesoscale Model (MM5). With respect to the ‘thermal’ part, the sensible heat storage flux is taken into account as a sink/source term in the thermodynamic equation at the surface layer. Shadowing effects are also considered as a function of the heat storage flux. The anthropogenic heat flux is introduced as a source term in the thermodynamic equation at the surface layer, based on the diurnal spatial distribution of the emission inventory. With respect to the ‘dynamical’ part, the surface stress and fluxes of heat and momentum under unstable conditions, as well as the diffusion coefficients under stable conditions, were modified, following recent advances over rough surfaces. The whole process was supplemented by detailed information on land use cover, derived from high resolution satellite image analysis. The model results were compared with sonic anemometer measurements of turbulence, tethered balloon soundings and routine meteorological data in Attica peninsula. The improvements, seen with the modified model, refer to the: (i) decrease of turbulence and fluxes during daytime (ii) decrease of the temperature amplitude wave (iii) strengthening of the nocturnal urban heat island (iv) the reduction/increase of the diffusion coefficient and potential temperature profiles during daytime/nighttime respectively (v) reasonable frictional retard in the sea-breeze front during daytime and the increase of the wind speed during nighttime (vi) estimation of the mixing height through the diffusion coefficients profile, while the differences in the calculated atmospheric boundary-layers depths are due to the lack of a unique definition application.