The main goal of this work was assessing the potential application of hyperspectral data to estimate the total suspended matter concentration in Barra Bonita Reservoir, Sao Paulo. The use of remote sensing data is a powerful motoring tool in water quality due to its capability to produce temporal and spatial data. Traditional monitoring techniques do not produce spatial and temporal information properly, and then, remote sensing data can help to overcome this gap. In order to estimate water quality parameters, quantitatively, the bio-optical models can be used to correlate aquatic compounds and water optical spectral properties registered by remote sensors. In situ hyperspectral measurements were used to calibrate and validate the bio-optical models that have already been published, and to create a new empirical model, based on the correlation between the total suspended matter concentration and remote sensing refl ectance values. Obtained results were evaluated by using error analysis (Root Mean Squared error - RMSE and Mean Absolute Percentage Error – MAPE, and coeffi cient of determination - R²), as well as the use of spectral bands from orbital sensors were also analyzed. The best TSM retrieval model used near infrared region, presented the slightest mistake to estimate suspended solid concentration.