This paper describes the mapping of vegetation patterns in two southern Mongolian protectedareas, namely the Gobi Gurvan Saykhan National Park and the Great Gobi B Strictly ProtectedArea. Both areas were mapped individually. Field-sampling followeda modified Braun-Blanquetapproach, relev ́es were classified with phytosociological methods. Training areas were assignedto phytosociological units, and were enlarged using various algorithms such as red-green-blueimages, NDVIs, Tasselled Cap, and unsupervised classifications. Theenlarged training regionswere used as ground truth data for a supervised classification of Landsat ETM+ scenes. Theclassifications were performed with a maximum likelihood algorithm and made more homogenousby applying a nearest neighbor filter with a 7 x 7 pixel range.Accuracy was checked with additional reference data sets taken in the field. The overallaccuracy was well above 90 %. Misclassifications occurred mainly due tothe ecotonal characterof some plant communities, or due to the occurrence of mixed pixels.In order to make resultsaccessible for park management and administration, the final data sets were implemented in aGIS that includes other thematic information such as country and districts boundaries, wells,rivers, lakes, as well as a high resolution digital elevation model.The mapped areas represent two out of five protected areas in Mongolia that are situated inthe southern Gobi. In this paper, they are for the first time compared with respect to theircommunity composition and their vegetation distribution