Abstract The aim of this study is to test, in ambient conditions, the capacities of semiconductor gas sensors to discriminate cured products (dry sausages of various origins or cured hams of different qualities) as well as aromatic or pathogenic bacterial strains commonly used for curing or sometimes encountered in meat products. An overall analysis of the dynamic headspace desorbed by the different substrates analysed has been carried out by six gas sensors with different characteristics. Using factorial discriminant analysis, we classify 94% of dry sausages, 87% of cured hams and 98% of bacterial strains into their respective groups. Cross-validation of the discriminant functions confirms the stability of the results. The results of this study show that with a limited number of semiconductor gas sensors, it is possible to classify rapidly and reliably various products, on the basis of the volatile compounds which they desorb.