The core of the largest Korgalzhyn State Nature Reserve in the Republic of Kazakhstan is the wetlands, which occupy an area of about 2.6 thousand km² and represent an alternation of numerous salt and fresh shallow lakes overgrown with aquatic vegetation, vast areas of shallow water, as well as terrestrial ecosystems of varying degrees of humidity with diverse vegetation. The reserve is the most important center for the conservation of biodiversity in Eurasia and is included in the Ramsar List of Wetlands of International Importance. During the moulting period and seasonal migrations, the number of birds can reach 5 thousand individuals. Wetlands are also a habitat for many species of animals and plants. One of the problems of the reserve is the extremely unstable hydrological regime, which leads to strong fluctuations in the area of the water surface of lakes and the degree of waterlogging of adjacent territories, and, consequently, to a change in the composition and spatial structure of plant communities. Changes in habitat conditions lead to fluctuations in the number of avifauna and depletion of species composition, which poses a threat to the biodiversity of the reserve. In this paper, based on remote sensing data, the change in the vegetation cover of the reserve wetlands is analyzed in connection with the change in the degree of waterlogging of the territory using the example of two years (low-water 2006 and high-water 2019). The paper describes a method for creating multi-temporal vegetation maps, which includes several successive stages: selection of terrestrial ecosystems as the main object for studying vegetation based on spectral indices; calculation and comparison of the degree of waterlogging of the territory for the years under consideration; selection and justification of the number of classes and the generalization coefficient of the results of unsupervised classification of images for detailing and clarifying the boundaries of plant communities; solution of the problem of thematic interpretation of the obtained spectral classes using the vegetation map of the 2006 GEF/UNDP project and field data; creation of multi-temporal vector maps, calculation and comparison of the area of the identified plant communities. The proposed approach allowed us to significantly detail the contour part of the vegetation map of the 2006 GEF/UNDP project and to carry out a more detailed division of plant communities. The calculations and comparison of the situation in 2006 and 2019 showed that the ratio of areas with different degrees of waterlogging can vary by 2–3 times depending on the year. The results of vegetation mapping confirmed the variability and dependence of the composition and spatial structure of plant communities primarily on the degree of waterlogging of the territory.