Playas in endorheic basins are of environmental value and highly scientific because of their natural habitats of a wide variety of species and indicators for climatic changes and tectonic activities within continents. Remote sensing, due to its capability of acquiring repetitive data with synoptic coverage, provides a unique tool to monitor and collect spatial information about playas. Most studies have concentrated on evaporite mineral distribution using remote sensing techniques but research about grain size distribution and geomorphologic changes in playas has been rarely reported. We analysed playa morphodynamics using Landsat time series data in a semi-arid endorheic basin, Salar de Uyuni in Bolivia. The spectral libraries explaining the relationship between surface reflectance and surficial materials are extracted from the Landsat image on 11 November 2012, the collected samples in the area and the precipitation data. Such spectral libraries are then applied to the classification of the other Landsat images from 1985–2011 using maximum likelihood classifier. Four types of surficial materials on the playa are identified: salty surface, silt-rich surface, clay-rich surface and pure salt. The silt-rich surface is related to crevasse splays and river banks while the clay-rich surface is associated with floodplain and channel depressions. The classification results show that the silt-rich surface tends to have a positive relationship with annual precipitation, whereas the salty surface negatively correlates with annual precipitation and there is no correlation between clay-rich surface and annual precipitation. Salty surfaces seem to consist primarily of clay due to their similar characteristics in response to precipitation changes. The classification results also show the development of a crevasse splay and avulsions. The results demonstrate the potential of Landsat imagery to determine the grain size and sedimentary facies distribution on playas in endorheic basins.