Remote sensing ground feature extraction is a key to remote sensing applications. AS the Remote sensing applications advance

various thematic information extraction methods and models have been researched upon and developed. In recent years

intellectual information extraction has become the focus of remote sensing specialists

as artificial intelligence advances in the computer field. The main objective of this paper is to develop a new method for extracting remote sensing information intellectually. Firstly

a knowledge-based model for thematic information extraction will be developed on the basis of remote sensing information mechanism study; Secondly

the model has been set up under the direction of OO idea and so a broad-sense knowledge-based method is used. The above intellectual information extraction method is illustrated with an example of water recognition automatically from NOAA/AVHRR. Satisfied results have been achieved when this method is applied to identify natural water and newly flooded areas. Finally

flood disaster monitoring has also been experimented in this way.