Turbidity is an important quality parameter of water from its optical property point of view. It varies spatio-temporally over large waterbodies and its well distributed measurement on field is tedious and time consuming. Generally, normalized difference turbidity index (NDTI), or band ratio, or regression analysis between turbidity concentration and band reflectance, approaches have been adapted to retrieve turbidity using multispectral remote sensing data. These techniques usually provide qualitative rather than quantitative estimates of turbidity. However, in the present study, spectral similarity analysis, between the spectral characteristics of spaceborne hyperspectral remote sensing data and spectral library generated on field, was carried out to quantify turbidity in the part of Chilika Lake, Odisha, India. Spatial spectral contextual image analysis, spectral angle mapper (SAM) technique was evaluated for the same. The SAM spectral matching technique has been widely used in geological application (mineral mapping), however, the application of this kind of techniques is limited in water quality studies due to non-availability of reference spectral libraries. A spectral library was generated on field for the different concentrations of turbidity using well calibrated instruments like field spectro-radiometer, turbidity meter and hand held global positioning system. The field spectra were classified into 7 classes of turbidity concentration as <5, 5–10, 10–15, 15–25, 25–45, 45–100 and >100NTU for analysis. Analysis reveal that at each location in the lake under consideration, the field spectra matched with the image spectra with SAM score of 0.8 and more. The observed turbidity at each location was also very much falling in the estimated turbidity class range. It was observed that the spectral similarity approach provides more quantitative estimate of turbidity as compared to NDTI.