Land cover change detection is an important applica ti on field of remotely sensed data In this paper

a new method for vegetation chan ge detection based on multitemporal Landsat TM images is introduced. Results of its application in vegetation change detection of Dunhuang Qasis show positive a spects of this method Color composition of multitemporal TM data gives an obviou s view of vegetation change in the stady area. Color components from color space transformation can be used for quantitative analysis of vegetation change. Hue and Saturation components indicate types and intensity of changes respectively. Lightness component shows general NDVI distribution over the study period. This new method can be easily applied to other change study fields.