Droughts and climate fluctuations can lead to seasonal drying in Etosha Lake, located in northern Namibia. Repetitive rises in temperature and lack of precipitation affect the hydrology and ecosystem health of using landscape of the Etosha Pan. Land cover dynamics of this salt ephemeral basin, located in Namibia, are subject to the climate and meteorological setting. To date, the spatiotemporal monitoring of this specific region of southern Afri-ca, including the driving factors of salinity and the water cycle, and the drainage dynamics of the lake, remains unclear. The remote location of this area and the extreme desert climate make fieldwork in this region a challenge. Using a series of six multi-spectral Landsat 8-9 OLI/TIRS satellite images and cartographic products (CORINE and GEBCO for thematic and topographic mapping), we identify seasonal variations in the surface of the Etosha National Park affecting drainage events in the lake basin. Extreme heat periods (summer-early autumn) resulted in the drying of the basin, which was covered by the crust of salt and minerals, while wet periods in winter and early spring favour the growth of vegetation. Technically, this paper presents the use of the Machine Learning (ML) methods of GRASS GIS by libraries of Python Scikit-Learn for image classification by an ensemble learning approach with a Random Forest (RF) classifier. Land cover types were identified using ML modules of GRASS GIS and scripting techniques. The methodology of scripts is presented in the GitHub repository of the author. The results demonstrated seasonal landscape dynamics in Etosha Pan. The ML method of image classification proved to be an effective tool for monitoring changes in the landscapes of northern Namibia, Africa.