A description and classification of the characteristics of space are the fundamental factors in building a multidimensional model of the natural environment. One of the criteria of describing an environment is analytical inferring and evaluation of the dynamics of phenomena based on structural attributes. The structure of space depends on the point of view. In this study, also in landscape description, were used fractal dimensions whose calculation algorithms were based on the „triangular prism method” in both the global and local options. To test the method, scanned infrared spectrozonal aerial images of the shore zones of the Mazury lakes Mikołajskie, Śniardwy and Łuknajno were chosen. There are areas with variously developed littoral zone embracing belts of marsh vegetation, emerged plants (reeds), floating plants, as well as the – also visualized in the images – zone of submerged plants growing on the bottom of the water body. To compare the calculated fractal dimensions, fragments were chosen of the mentioned water bodies representing different zones of shoreline vegetation, as well as the surrounding land under agricultural use, forest cover and urbanized areas. The results of the investigation revealed that, with appropriate standardization of remote sensing data preparation, the fractal dimension can be a valuable indicator of quality of landscape diversity.