In addition to the traditionally complex land surface emissivity models, it is possible to direct real-time retrieve emissivity over land surface in a large-scaled temporal and spatial region from satellite microwave observations. We investigated the performance of the real-time dynamic retrieval emissivity against the monthly averaged emissivity atlas. As indicated by the results, the dynamic retrieval algorithm can obviously reduce the first guess departure of low-level channels which are sensitive to the surface and the simulated background brightness temperatures with real-time dynamic emissivity against the observations have a better agreement over the land surface and higher correlation. Therefore, the dynamic retrieval emissivity algorithm has the potential to improve the effect of satellite data assimilation.