The complexity of urban climatology is mainly due to the surface heterogeneity of, e.g., roughness, land use, etc. Thus the urban climatology is composed of many microclimates in the various urban quarters. Mesoscale models integrate the surface heterogeneities within each grid cell: it thus appears important to evaluate the influence of the representation of canopy and soil in urban areas in the simulations of the Urban Boundary Layer (UBL). Here, the urban soil model, SM2-U, is coupled with the French communal model SUBMESO. This model, SUBMESO, initially derived from ARPS 3.0, is mainly developed for simulating the atmospheric boundary layer with very high resolution. SM2-U is an extension to the urban surfaces of the rural soil model SM2-ISBA previously derived from the force-restore model of Noilhan & Planton (1989). SM2-U considers 5 cover modes: natural soil, bare soil, artificial soil, building roofs, and water surfaces. Three soil layers are considered, a superficial layer for the non urbanized surfaces, a second soil layer representing the influence zone of the vegetation roots, and a third layer used as a water reservoir for the second layer during the dryness period. In each computational cell, the model determines a surface temperature and a specific humidity of the apparent surface for each of the 5 cover modes, and computes their average according to the cover fraction. The urban canopy influence is considered within the surface temperature equation of the artificial soil by introducing the heat storage by buildings walls, and by taking into account the radiative trapping with a street canyon effective albedo parameterization deduced from the work of Masson (2000). Coupling SUBMESO and SM2-U is tested above a hypothetic city, with a flat ground and infinite lateral extension (periodic boundary conditions). Simulations are carried out on a diurnal cycle starting at midnight. Dynamics and thermodynamics inside the computational domain are forced only by the ground fluxes. Three simulations are compared. In the first one, the city is composed of 4 quarters: residential, high rise buildings, city center, and industrial-commercial quarter. In the second simulation, the city is represented by a single quarter averaging the four preceding quarters. In the third one, the waterproof surfaces in the first simulation are replaced by a dry bare soil. The results of the first simulation illustrate the influence of each urban quarter on the ABL structure, and the generation of heat island effect, whereas the case of the averaged city shows a decrease of the urban area impact on the ABL structure, and a decrease of the heat island intensity. The results of the third simulation show the difficulty to reproduce the urban effects by using only the rural parts of SM2-U.