The goal of this study is to examine the impact of assimilating satellite derived surface temperature over land (LST) in the surface scheme of AROME-France model. The LST is retrieved from SEVIRI radiances during the assimilation process in the atmospheric model. The assimilation of LST is performed using an optimal interpolation technique, similarly to the assimilation of other near-surface parameters (temperature and relative humidity at 2 meters). Observation and background errors were diagnosed before to prescribe them in the surface assimilation scheme. First, this LST assimilation has been evaluated in terms of analysis and forecast quality over a two-month summer period. A positive impact has been found on the assimilation of 2 m temperature and relative humidity with a slight decrease in bias of the background departure. An improvement has also been found for the assimilation of microwave humidity sensitive channels. The assimilation of microwave sensors benefits from an updated land surface temperature through the retrieval of the emissivity. Moreover, the assimilation of SEVIRI LST has improved the nighttime forecasts of temperature and relative humidity near the surface and up to 700 hPa. Several open issues for improving these preliminary results are finally proposed.