The emphasis of this research is to demonstrate application of Landsat satellite imagery as a major resource for student and educational research. Landsat images are highly useful and strongly recommended for educational purposes as they are provided free of charge and timely regular geospatial data with 30-m resolution covering any places of the Earth. The case study describes mapping land cover types in ecosystems. It details how exactly satellite images can be used for geospatial research step by step. The research methodology is based spatial analysis tools of the open source GIS software: Quantum GIS and ILWIS GIS. The pre-processing of the imagery included data import and settings of colors and contrast. The contrast of the Landsat TM imagery has been enhanced using «Stretch To MinMax» function in QGIS viewer. The series of Landsat TM images, which cover Bovanenkovo region in Yamal peninsula with 30-m resolution, were initially stored as raster file in Erdas Imagine (.img) format. The images were imported to ILWIS from .img into ILWIS .mpr raster map format (ASCII) using GDAL. After converting, each image contained collection of 7 Landsat raster bands, as well as theirs metadata stored in Map List (.mpl) file, information about georeference (.grf) and coordinate system in .csy file. The coordinate systems of the Landsat imagery is based on the WGS 1984 datum. To improve the quality of image and visibility of coastal lines, the image enhancement of the Landsat bands was performed using Edge Enhancement linear filter from the Image Processing menu. Stretching filter was applied to the bands, in order to improve color contrast: since not all pixel values have been used in the image, the actual diapason range was stretched linearly to 0-255 values. To visualize spectral information from the Landsat image, a color composite map has been created using combination of three raster images of the individual bands. Supervised classification of the raster imagery includes image analysis aimed at recognizing class membership for each pixel. The main research method used in current research is supervised classification, which enabled to assign land cover classes by adjusting classification parameters and thresholds in DN values of spectral signature of pixels. The principle of Minimum Distance method, used for pixels classification is based on the calculating of shortest straight-line distance in Euclidian coordinate system from each pixel’s DN to the pattern pixels of land cover classes. The researched described successful application of Landsat remote sensing data and GIS for environmental studies.