A semianalytical approach to the problem of determining inherent optical properties from satellite and in situ ocean color data is presented. The model uses empirically derived spectral slopes between neighboring wavebands in combination with radiative transfer modeling to determine the spectral absorption (alpha) and backscatter (b(b)); these values are then further decomposed into absorption due to phytoplankton, detrital, and colored dissolved organic matter components. When compared with over 400 in situ data points the model makes good retrievals of the total absorption and backscatter across the entire spectrum, with regression slopes close to unity, little or no bias, high percentage of variance explained, and low rms errors.