The water quality of the Tat River, referred to nutrients compounds, is lower than the expected. Its management must be largely improved. The present work takes part in a global effort of development and evaluation of reliable and robust tools, with the aim of allowing the control and supervision of its lowland area (at the south Mediterranean coast of France). A simplified model, based on mass balances, has been developed to estimate nitrogen and organic matter concentrations in the stream and to describe the river water quality. Kohonen self-organizing traps (KSOMs) were used to deal with missing data. This kind of neural networks proved to be very useful to predict missing components and to complete the available database, describing the chemical quality of the river and the WasteWater Treatment Plant (WWTP) outflows. The simulation model also proved to be a good tool for the system evaluated. The results it provided reveal the high impact of the WWTPs located along the studied area, due to malfunction and tourism activities. (C) 2007 Elsevier B.V. All rights reserved.