The retrieval of aerosol properties using multi-spectral data provided by satellite based radiometers, has been explored by TNO since 1995. Aerosol properties that are currently retrieved using (A)ATSR data are aerosol optical depth (AOD), the variation of AOD with the wavelength expressed by the Angstrom coefficient, and the dominant aerosol types. Algorithms have been developed for instruments such as AVHRR, GOME, SCHIAMACHY, OMI and (A)ATSR. A major challenge is to discriminate between the aerosol reflectances at the top of the atmosphere (TOA) and those by other atmospheric constituents and, in particular, the surface reflectance. Over water the surface is usually dark and the TOA reflectance is dominated by contributions from atmospheric constituents. Hence only one single viewing angle suffices for the retrieval of aerosol properties over water. The accuracy obtained for AOD retrieval over water using ATSR-2 data is 0.04, based on comparison with independent ground-based measurements, usually AERONET sun photometer data. Coastal waters often cause a problem because of sub-surface reflectances by suspended matter and the retrieval is often less reliable when this occurs. Also the presence of algae blooms (Chlorophyll) may render less accurate results. This causes discontinuities in the AOD across land-sea boundaries. Land surfaces are usually brighter and often the contribution of the surface to the reflectance at the TOA is significant and needs to be accounted for to accurately determine the aerosol properties. The surface contribution to the TOA reflectance can be eliminated by using multiple viewing angles such as provided by (A)ATSR. (A)ATSR provides one viewing angle in nadir and one in the forward direction. This feature has been applied over land in various regions around the world and comparisons with AERONET data provide a means to evaluate the results, showing an excellent accuracy of 0.05-0.06 over most land surfaces. The TNO single and dual view algorithms are scientific algorithms, which have been used to develop a quasi-operational algorithm that has been demonstrated to work well over Europe. Data over Europe are available as a service through GSE PROMOTE and TEMIS. The application for 2003 is in progress. This quasi-operational algorithm will be expanded to other areas such as Asia and Africa, including deserts, based on scientific work in progress. The presentation will focus on the application of the quasi-operational algorithm over land, and scientific work on its extension to other areas than Europe.