This present paper presents a new inversion method of inland water based on spectral matching. First step of this method is using the known water surface measured hyperspectral and the absorption coefficient of each component, and obtaining backscattering coefficient of suspended matter with bio-optical model. The second step is calculation of the spectral reflectance of water bodies based on bio-optical model, through the cross-combination of inherent optical property of water components (chlorophyll, suspended matter, yellow substance), in order to create a look-up table of Rrs that corresponds to all combinations of water component. The third step is changing the look-up table data into MODIS spectral data using MODIS channel response function. The final study examined the applicability of the look-up table using the Hyperspectral and MODIS bands spectra, based on the minimum distance principle, to find the best matched spectra, thus it has found corresponding concentrations of three components. The average relative error of chlorophyll and suspended matter is 38.6% and 28% respectively. Optical properties of water components interfere with each other because of the complexity of inland water bodies, resulting in difficulty to extract the feature band for statistical model, and while it is hard for bio-optical model inversion method to solve the unstable problem of special inherent optical property with the seasons and regional issues, the method of this paper would have the advantages of bio-optical model, while eliminating instability of special inherent optical property, so it is a good approach to inland water retrieval.