The experimentation takes place in the Maya Biosphere Reserve, in the heart of the Peten region in Guatemala. In this natural area intermingled rivers and lakes, the forest which was in balance with environmental conditions dominated all the space. However, the landscape has just suffered a real transformation for the last 15 years. Since 1987, populating has settle up regularly by succesive waves. They have appropriated, cleared and changed the native forest in pasture and milpa (field of corn). This process of systematic deforestation by large fires, permits the creation of new rural societies, a new area of distinctly diverse uses. But the sudden and non control setting up of these populations threaten environment conditions. A conflict for the land has been appeared around and inside of the Maya Biosphere Reserve, whitch is itself threatened. The State of Guatemala, as the NGO need a local and regional perception. And yet, faced with this speedy phenomenon non finished, the lack of updated cartographic data in a area little known and badly statistical informed, high resolution remote sensing becomes an irreplaceable tool to understand such radical transformations. To understand spatio temporal process of this new rural pioneer front, to make a dynamic diagnosis, to date, to follow, to map, to update environmental and statistical data, the method of image processing proposed is based on satellite data fusions--Landsat-TM and Spot--by multidated approaches (11 images over 15 years), multi-scale (from local to regional) and multispectral (only one image resultant of 41 georeferenced channels) ; the results have been ratified by field work.