Efficient methods of non-destructively measuring chlorophyll content (Chl) across a wide range of greenness in grape leaves are necessary for a producer to effectively monitor the status of the crop. Wet chemical methods and handheld leaf Chl meters such as the Minolta SPAD-502 Chl meter(Markwell,1995) have traditionally been used for this purpose. Recently, the availability of small and affordable radiometers , which uses Near infra-red and Red edge(CIred edge)( Mark R. Steel, 2008b), for Chl estimation in grapevine leaves, has provided means to estimate Chl from reflectance measurements. The two methods were equally accurate measuring pigments at low to moderate Chl levels. But, when Chl exceeded 300 mg/m2, SPAD sensitivity to Chl declined noticeably and the accuracy of Chl estimation by the CIred edge was much higher than that of the SPAD meter, the chlorophyll index was found to be capable of accurately estimating pigment contents across a much greater Chl range than the SPAD meter. However, in this paper an artificial-intelligence technique, the Support Vector Machine (SVMLightV6.01) model was introduced to establish the relationship between the Chlorophylls content and reflectance of 400-750 nm spectrum, variation of species and growth stages, which can much more perfectly compensated the reflectance that was absorpted by anthocyanin and carotenoids and scatterd because of different leaf thickness, density, or surface properties. As a result that we have solved the problems such as saturation and/or asymptote in high/or lower content of chlorophyll, and SVM model was found to be capable of more accurately estimating pigment contents across a greater range than the SPAD and CIred edge, thus it can be used for quantitative assessment of early stages of pant stress.