The use of the chemometric techniques to exploit the selectivity of gas sensors, although current in the case of multisensors systems, is tested here with only one gas sensor. The method consists in measuring the electric conductance of the sensitive element, a tin dioxide bar, at various temperatures. Although these gas sensors follow non linear laws, the modelization algorithm Partial Least Square (PLS) is preffered to neurak networks. A comparison of the two algorithms is made, as well as a study of the prediction. Lastly, theconcept is tested on an autonomous prototype in real conditions.