In recent years, air pollution has been a subject of great scientific and public interests for the strong impact on human health. Air pollution is due to the presence in the atmosphere of polluting substances, such as carbon monoxide, sulfur and nitrogen oxides, particulates and volatile organic compounds (VOCs), derived predominantly from various combustion processes. Benzene is a VOC belonging to group-I carcinogens with a toxicity widely demonstrated. The emission limit values and the daily exposure time to benzene (TLV-TWA) are 5μg/m3 (0.00157 ppm) and 1.6mg/m3 (0.5 ppm), respectively. Currently, expensive and time-consuming analytical methods are used for detection of benzene. These methods require to perform a few preliminary steps such as sampling, and matrices pre-treatments. In addition, it is also needed the support of specialized personnel. Recently, single-walled carbon nanotube (SWNTs) gas sensors with a limit detection (LOD) of 20 ppm were developed for benzene detection. Other innovative bioassay, called bio-report systems, were proposed. They use a whole cell (Pseudomona putida or Escherichia coli) as molecular recognition element and exhibit a LOD of about 10 μM. Here, we report on the design of a highly sensitive fluorescence assay for monitoring atmospheric level of benzene. For this purpose, we used as molecular recognition element the porcine odorant-binding protein (pOBP). 1-Aminoanthracene was selected as extrinsic fluorescence probe for designing a competitive fluorescence resonance energy transfer (FRET) assay for benzene detection. The detection limit of our assay was 3.9μg/m3, a value lower than the actual emission limit value of benzene as regulated by European law.