Climate change influences crop production. Crop management will have to adapt the crop to this new climate. Models can be used to guide this new crop management. Using expert knowledge and surveys we developed a model that aims to simulate the number of days suitable for sowing maize. Our model is based on three modules: a module dealing with the sowing constraint due to frost; a module for constraints due to temperature to grow crop and a module for constraints due to soil moisture. The model was calibrated using recorded sowing dates in 2000 and 2003 and evaluated using an independent set of recorded dates in 2005 in southwestern France. Results are good, as only four farmers over 39 sowed on an apparently unsuitable date. Inaccuracy of input data such as rainfall is discussed and some modifications to the present algorithm are proposed.