Remote sensing of inland waters relies on the retrieval of optically active constituent concentration using reflectance as input to different types of algorithms. Global carbon cycle, sediment budgets, phytoplankton primary production and water quality are among processes that can be evaluated using remote sensing imagery. Thus, Sentinel-2 MSI (Multispectral Instrument) launch increased the possibilities for mapping and monitoring aquatic environments due to high spectral, spatial and radiometric resolutions. This work tested six established algorithms for estimating absorption by colored dissolved organic matter and concentration of total suspended solids and chlorophyll-a in an Amazonian floodplain lake (Curuai). Fieldwork data was used to simulate the MSI reflectance and to adjust regression models. Based on these models, a MSI image was applied to spatialize optically active constituent distribution over Curuai lake. Small range of constituent concentration and low signal level represent a huge challenge for CDOM retrieval in Amazon turbid waters, as shown by low determination coefficient (< 0.45) and high relative error (> 10%) provided by models. The adjustment of chlorophyll model showed a high correlation between in-situ and satellite observations (R² > 0.86), although larger errors were assessed in low chlorophyll concentration. Results were more robust for TSS retrieval, as expected in very turbid waters with wide range of concentration values. Lower accuracy was observed when models were applied to MSI image due to higher remote sensing reflectance values, therefore resulting in an overestimation of TSS and Chl-a concentration.