Change vector analysis and principal components analysis in NDVI time trajectories space are powerful tools to analyse the land cover change. The length of the change vector indicates the magnitude of the change

while its direction indicates the nature of the change. This change detection method is applied to two remotely sensed indicators of land surface conditions

i.e. NDVI and spatial structure

with aims to improving the capability to detect and categorize land cover change. The magnitude and type of changes and the changes in spatial structure indicator in China from 1983 to 1992 are calculated using the above mentioned analysis methods. Through the research

the main conclusions are drawn as follows:\;1) The change vector magnitude of NDVI is different apparently between East and West China. Larger changes of NDVI lie to the Southeast coastal areas

Taiwan and Sichuan

Yunnan province （monsoon area of Southeast and Southwest China）. The changes in Northwest area and Qinghai Xizang Plateau are little. It can conclude that the monsoon climate and the human activity impact apparently on the land cover dynamic of East area （especially the different kinds of crops）.\;2) NDVI of land cover changes is derived from many factors （climate

hydrology

landform

soil

land use etc.）. Although the basic tendency of the climate in North China is that the temperature becomes increasing and precipitation get decreasing （drying）

no obvious changes are presented in large scale in this region. Landform types and vegetation types in North China are the key factors for NDVI changes.\;3) The spatial structure indicator is able to detect changes in the seasonal ecosystem dynamic for spatially heterogeneous landscapes. Changes of spatial structure indicators of NDVI mainly locate in South China

reflecting the high heterogeneity vegetation during the growth and decay processes. The areas composed of Tianshan Mountains

Changbai Mountains

and Xiaoxinganling Mountains reflect the impact of landform types on the spatial structure index.