This paper discusses the relationship between the values of chosen landscape metrics and the generalisation levels of vector as well as of raster vegetation maps. It is shown that parallelly to progressing generalisation, the distribution of the particular patch types changes. After the seven repetitions of the generalisation procedure (with minimum patch (grid cell) size increasing 64 times) one could distinguish four main association groups, differing as to the nature of changes in spatial shares. The first encompasses types of patches, whose spatial shares decreased on both maps analysed. The second - these types, whose area shares decreased on the vector map, and increased on the raster map, The third-types, whose spatial share on the vector map increased, and decreased (sometimes even down to disappearance) on the raster map. The fourth group contains the vegetation types, which increased their spatial share on both maps with simultaneous increase of the mean patch size. It is also shown that values of the following indices: total edge (TE), edge density (ED), total number of patches (NUMP), patch size coefficient of variance (PSCOV), number of patch types (RICH), area weighted mean patch fractal dimension (AWMPFD) and area weighted mean shape index (AWMSI) decrease along with the generalization level. The next group of indices include mean patch fractal dimension (MPFD) as well as mean shape index (MSI). These indices display the entirely different directions of changes on the vector and raster maps. On vector maps the value of indices initially increases, and then, for the basic field equal to or bigger than 16 000 sq.m, it remains at virtually constant level. On raster maps the value of indices initially decreases, but for the basic fields bigger than 16 000 sq. m it also remains at the constant level. For the last group: patch size standard deviation (PSSD), mean patch size (MPS), and mean patch edge (MPE), values of indices increase with the degree of generalisation, and the increase is steeper for the vector maps than for the raster maps.