We report on a simple microsystem for the analytical quantification of benzene in parts per billion (ppb) concentration. The system is based on a commercial photoionization detector (PID) and a MEMS cartridge, filled with innovative supramolecular cavitand receptors. The heater integrated on the MEMS device enables fine tuning of its temperature and operating the cartridge as both purge-and-trap and gas chromatographic (GC) column. By means of a smart signal process algorithm based on fuzzy neural network (FNN) the system is able to exactly quantify benzene also in mixture containing others aromatic species. Functional characterization results are shown.