A WAY TO increase crop productivity is to use models and decision support tools. The Tsigane device offers a geographic information system including the crop growth model, MOSICAS, and georeferenced data required for the simulations (agricultural, meteorological, geographic data). Different solutions have been developed to feed the database: an automatic weather stations network, an administrative organisation to monitor the location of agricultural land parcels and corresponding acreage, and user interfaces to input agricultural data. The system makes it possible for the users to consult the data in a cartographic form using queries in the database, to be warned in case of climatic risks (drought, excess of water, etc), and to run crop growth simulations and yield predictions. The yield prediction errors for the 2006 harvest were lower than 5%. We also produced potential production maps of different areas under various scenarios (different climatic type, different harvest dates) with the data of each land parcel of Guadeloupe island. This system is a very powerful tool for the management of sugarcane production. It is open to everyone thanks to internet technology. Farmers can optimise their agricultural practices; government services get objective and reliable data on the sugarcane land parcels and may anticipate more accurately the sugar production and the required public fmancial aids; agricultural technical centres can provide farmers relevant cropping recommendations and co-ordinate more efficiently the planting and harvesting operations over a whole region.