Taking the core area of the Three Parallel Rivers Belt in Gongshan county of Yunnan province as the example, the distribution of the landscape types was researched by GIS, RS and GPS technology. The results showed that: (1) The matrix of landscape was forest vegetation, especially the dominant forest types along altitude at the alpine area. Because of the vertical differentiation of the landscapes distribution, the dominance value of the landscape types was not high. But along the altitude it could be found that mountainous humid evergreen broadleaved forest was the dominant landscape at the bottom of the valley area,the Tsuga mixed forest with evergreen broadleaved species at the mid -mountain and sub -alpine area, the coniferous forest dominated by Picea, Abies and Larix at the sub-alpine and alpine area, and Rhododendron shrub at the alpine area. (2) For dominant landscape types, their characteristics were larger area, large number of patches, large average patch area, big largest patch index, high core area and core area index, long edge length, big edge density, big shape index, big landscape fractal dimension, and patches distribution was relatively concentrated. (3) For zone vegetation landscape types and the vegetation landscape types which were not zone vegetation located in the vertical belt, their characteristics reflected small area of patches, small number of patches,mean patch area relatively in medium, small largest patch index, small core area, relatively few patches number of core area. The shape index was big, and the landscape fractal dimension was smal造, and the cohension index, MESH and the aggregation index was small with the landscape. (4) For different artificial patches, the area features were different.(5) The patterns of non-vegetation landscape types were different because of their different distribution areas. (6) The resident settlements showed the characteristics of small area, small number of patches, small core area, low shape index and fractal dimension, and the patches were scattered. |