It is a key problem to analyze quantitatively the correlations between remote sensing data and chlorophyll-a comcentration for measuring the chlorophyll-a content with remote sensing method and technology. Two field campaigns were made in March and June 2003

respectively. Reflectance spectra were measured with a portable FieldSpec FR spectroradiometer （ASD Inc.）. Water samples for chlorophyll-a measurement were collected at the same time guided by some related standards of China. The features of the spectral reflectance were analyzed in detail with peak height

trough depth

area surrounded by peak

distance from peak to trough and peak shoulder width. And the same result as others is to discover the most important features near 682nm and near 706nm. The regressive linear/nonlinear fitting model were developed respectively between chlorophyll-a concentration and other factors as follows: 1st

the ratio of reflectance near 706nm to reflectance near 682nm （R706/R682）;2nd

the ratio of reflectance near 706nm to reflectance near 572nm （R706/R572）;3rd

the first derivative of reflectance near 690nm （R’694）;4th

the position of reflectance peak near 706nm（λmax706）;5th

the distance from reflectance peak near 706nm to trough near 682nm （Dpeak706-682）;6th

the area surrounded by reflectance peak near 706nm （Apeak）; and 7th

the height of reflectance peak near 706nm （Hpeak706）. By comparison of R2

average error and RMS error respectively

the results show that R706/R682

exponential fitting degree R2 with 0.9199

λmax706

linear fitting degree R2 with 0.9038 can denote chlorophyll-a content and they are better to predict chlorophyll-a concentration of water-body near Meiliang bayou in Taihu. The position of reflectance peak near 706nm is the most sensitive factor to denote chlorophyll-a concentration by comparison of distance from reflectance peak to absorbance value

height of reflectance peak and area of reflectance peak. It is very important to predict chlorophyll-a content for the bands near 572nm

682nm and 706nm.