Climate change and demographic growth of the world population force farmers and the agricultural sector to adapt in order to face these two major challenges. Cultivated on almost 20% of the total land area, croplands largely contribute to global greenhouse gas emissions. These areas also offer the possibility of implementing climate change mitigation strategies. In this context, this thesis work aims to improve our knowledge on the functioning of croplands and to provide tools to assess the contribution of cultivated lands to climate change, and to quantify biogeochemical (C storage) and biogeophysical (albedo effect) contributions to climate change mitigation through cover crop implementation. To meet these objectives, two modelling approaches were developed during my PhD. The first part of this thesis aimed at developing a spatialized modeling approach able to provide estimates of production (biomass, yield), CO2 and water fluxes, as well as carbon and water budgets for croplands. To this end, the SAFYE-CO2 agrometeorological model, which assimilates remote sensing products of vegetation index (green area index) with high spatial and temporal resolutions, was developed and applied to several crops (wheat, maize, sunflower) and intercropping vegetation (spontaneous regrowth, weeds, cover crops). This approach was validated on a network of plots located in southwest France, benefiting from a large number of satellite images and validation data collected over the Regional Spatial Observatory. Notably, it allowed to precisely estimate wheat, sunflower and maize productions, as well as CO2 and water fluxes on wheat and sunflower. The vegetation that developed on the plot during the fallow period was also taken into account in order to improve CO2 and water fluxes estimates. This leads to a better quantification of the cover crop impact on the carbon budget components of croplands in the study area. The second part of this thesis aimed at developing a model assessing the radiative forcing induced by surface albedo changes following the introduction of cover crops in the crop rotations at European scale. We took advantage of coarse-resolution remotely sensed albedo products (1/20°) developed by the CNRM (and in collaboration with this laboratory) to quantify this albedo effect. Several scenarios of introduction were simulated to quantify the effect of certain factors such as rain or snow cover. They allowed to alert on the potential strong negative impact of soil darkening (due to soil organic carbon increase) induced by cover crops on cropland radiative forcing at long term. Finally, as any change in agricultural management practices induced biogeochemical and biogeophysical effects on climate, a coupled analysis of these effects was carried out at a regional scale thanks to the combined use of these two modeling approaches. We concluded that once the cover crops are implemented, the soil should be permanently covered with vegetation so that soil darkening does not counterbalance the other climate benefits of this agricultural practice.