Traditional data model takes the geometric object as the modeling target and results directly in decomposing geographic phenomena into a series of objects. Furthermore

the temporal data and spatial data are stored separately. All of those restrict the development of GIS and resources monitoring. Feature based spatio temporal data model can support the next generation of GIS effectively. In this paper

based on the landuse monitoing research project of the five year plan of our country

we propose a new spatio temporal data model based on the states of the changing feature （SCFSTDM） in order to serve for integrating geographic phenomena and geographic information as well as applying object oriented analysis method to support spatio temporal analysis

simulation and data sharing. \;Finally

we develop a prototype of the temporal landuse information system in which there is a spatio temporal database based on the feature. With this system

we can query states of the feature instance directly

for example

to query the states and to play the developing process of a feature instance

to make the spatio temporal reasoning.