This paper provides a synergetic approach between numerical modeling and remote sensing of bio-optical water properties. The work demonstrates that appropriate data-assimilation schemes make numerical modeling a suitable and reliable tool for filling the gaps arising due to satellite imagery unavailability and/or cloud covering. In this research we apply the <i>Princeton Ocean Model</i> to the Sea of Azov, assimilating bio-optical indexes (<i>index</i>34 and <i>b<sub>bp</sub></i>(555)) from MODIS L2 products. These data identify the presence of suspended matter (mineral suspended matter from river discharges or resuspending as a result of a strong wind), and suspended matter of biological origin. The <i>ad hoc</i> assimilation/correction scheme allows for prediction (and reanalysis) of transport and diffusion of the bio-optical tracers. Results focus on the ability of the method to provide spatial maps that overcome the general issues related to Ocean Color imagery (e.g., cloud cover) and on the comparison between the assimilating and the non-assimilating runs. Methods of joined information analysis are discussed and the quality of model forecasts is estimated depending on the intervals of the satellite data assimilation. Hydrodynamic modeling of the Sea of Azov was carried out for the period of 2013–2014 applying meteorological data of the regional weather forecasting system <i>SKIRON/Eta</i>. The analysis of data coherence helps to detect negative changes to the sea waters, predict them and forecast typical areas and territories subject to anthropogenic impact. The successive data-assimilation algorithm is proved to improve the forecast of suspended matter transfer.