This study investigates the potential of pervaporation (a selective membrane separation technique) combined with an electronic nose based on metal oxide sensors for analyzing wine model solutions. Choosing a suitable membrane polymer, it is shown that subtle variations in aroma composition can be detected by the metal oxide sensors, even when 12% (v/v) ethanol is present in the original sample. Simulations of the composition of the permeate demonstrate that, despite the low molecular density of the permeate, the respective solute concentrations are sufficiently high to generate a reproducible and significant response from the metal oxide sensors.