Due to the fast social and economic development in its valley, the Shiyang River has experienced tremendous environmental changes, especially in the past decade. It is rather difficult to detect the long-period and complex changes of water quality merely by means of traditional measurements. In this paper, the remote sensing technology was used to monitor and evaluate the water quality in the Shiyang river. Combining the TMdata with the pollutant in the water, the authors developed an inversion model, with which one can easily evaluate the content of pollution in the water by using the TMdata. The historical Landsat (TM) data obtained in four consecutive years were employed to extract the changing information of the river water. The results show an alarming fast increasing trend of pollution in the river. The method was used to extract the water pollution information in the river, and the results show its good performance with a monitoring accuracy of 84%, suggesting a great potential of this method in this field .