Using the measured dry-leaf biochemistry and hyperspectral reflectance data,a new thinking clue of leaf biochemical retrieval is developed based on the 1st derivative extremum of area-nomalized hyperspectral position variables.Research results indicate that this method can be employed to effectively extract the concentration of foliar total nitrogen,cellulose,lignin and starch.Especially the retrieval precision of cellulose,lignin and starch contents is better than the research available.Moreover,the ultimate application direction of vegetation remote sensing is the its canopy level.The investigation suggests that this technique can effectively remove soil influence on the extraction of total nitrogen,cellulose and lignin content,but the retrival result of starch is still not satisfactory.