Abstract:The optimum vegetation indexand methods for collecting vegetation information in loess plateau area were studied and appraised. In this area, there were deciduous broadleaved forest, steppe and desert. Vegetation types were rich and diverse. However, collection of vegetation information was strongly influenced by the soil background. The modified soil-adjusted vegetation index (MSAVI) after being compared and analysed not only could enhance vegetation signal, but also could minimize greatly the soil background effects. It could also reflect the distribution of the vegetation types macroscopically in this area. At present, it has been a more or lessideal vegetation index for collection vegetation information in loess plateau area. The collection and monitering of vegetation information for the normalized difference vegetation index (NDVI), soil-adjusted vegetation index (SAVI) and perpendicular vegetation index (PVI) were used in high, medium density and rare vegetation area respectively. For a variety of vegetation indices, the images which were produced by cumulative multitemporal vegetation indices could extract vegetation information well reflecting the vegetation distribution of that particular area.