This paper provides a description of Kohonen’s self-organizing maps (SOM) and principles of original multidimensional data space transformation into a two-dimensional SOM. Also, features of ScanEx NeRIS software designed for analysis and visualization of satellite data with built-in SOM tool are presented. High-resolution images (in GeoTIFF format) obtained by Landsat (Landsat 5) radiometer are utilized as the input data. A NeRIS-based technique for water boundary line identification with satellite data for further qualitative flooding evaluation is elaborated. It is found out that class belonging probability threshold is a critical parameter of the proposed technique that affects the identification. Satellite data of flooded floodplain areas of the Upper Ob in June 2010 are processed. Vector maps of areas with data processed by SOMbased technique are obtained and exported to Quantum GIS for building a vector map of flooded areas. It is concluded that further quantitative analysis requires calibration of the identification technique parameters with ground data.