In order to improve the accuracy and speed and make large range monitoring and evaluation of lake eutrophication, this paper measured the spectrum and analyzed the water samples in the Baihua Lake of Guiyang City. By studying the normalization of water spectral reflectance, the first and second order differential spectrum processing, a diagnosis was made on the sensitive bands? of the various water quality parameters to establish the hyperspectral?quantitative model. This paper also discussed the quantitative relationship between the spectral reflectance and concentration of water quality parameter, evaluated the entrophication of water with comprehensive nutrition state index method (TLI). The results show that (1) it’s suitable to use hyperspectral remote sensing to monitor water quality, and the sensitive bands of the water quality parameters (Chl-a, TP, TN, SD and CODMn) are respectively 699nm, 823nm, 399nm, 563nm, 504nm; (2) The effect of the model with sensitive brand is very good, the various estimated models R2 are respectively 0.8276, 0.7711, 0.7917, 0.9213 and 0.8328; (3) The water quality is in moderate eutrophic state from the evaluation of eutrophication in the Baiha Lake based on the measured and estimated data.