A sensor array composed of selective and partially selective electrodes is applied to the recognition of beverages: milk, orange juice, and beer. The role of selective and partially selective sensors in classification results is discussed. For this reason two experiments are realized: distinguishing among various types of foodstuffs and the recognition of their origin (manufacturer). In both cases, the measurements were performed in stationary and flow mode. Classical ion-selective electrodes and miniaturized solid-state electrodes were considered. The best results are obtained when the two types of sensor are incorporated in the array, which leads to the highest correctness of the system response. It is however worth to notice, that in the case of some samples (for example, milk) satisfactory results were obtained when only selective or when only partially selective electrodes were applied.