The landscape of the Okavango Catchment is structured much more diverse than is perceptible at first sight. From the mountainous Angolan highlands with altitudes of more than 1700 meters above sea level to less than 1000 meters at the delta and the surrounding Kalahari sands; accordingly the relief intensity decreases steadily, from high mountainous to undulating in the center parts and flat in the south. Finding a way to describe, analyze and outline this diverse landscape structure and to classify landscape units on the basis of available data is just as challenging as useful. This paper shows how to derive geomorphographic units (GMUs) as discrete terrain entities on the basis of a SRTM digital elevation model and how to valorize and verify these units. The GMUs semi-automatically computed with SAGA GIS are based on a set of local and regional continuous land surface variables. As a first result we obtained 17 GMU-classes, sorted into four main groups: major valley floors (MVF), tributary valley floors (TVF), sandvelds and longitudinal dunes (SLD), and slopes and summit areas (SSA), scattered over the catchment. To test the applicability, we correlated the GMUs with two different vegetation indices (NDVI, EVI), a MODIS landform classification and soil parameters. In conclusion, single GMU-classes respond reasonable in a distinct and specific way to the land cover and reproduce the physiogeographic settings of the Okavango catchment quite well. On the basis of such units, further delineations respectively mappings of, for example, vegetation or soil data is seen as the next step. Keywords: DEM, geomorphographic units (GMUs), MODIS, semi-automated landscape analysis, SRTM.