Analysis of exhaled air leads to the development of fast accurate and non-invasive diagnostics. A comprehensive analysis of the entire range of volatile organic compounds (VOCs) in exhaled air samples will enable the identification of VOCs unique for certain patient groups. This study demonstrates proof of principle of our developed method tested on a smoking/non-smoking study population. Thermal desorption and gas chromatography coupled to time-of-flight mass spectrometry were used to analyse exhaled air samples. The VOC profiles obtained from each individual were combined into one final database based on similarity of mass spectra and retention indexes (RI), which offers the possibility for a reliable selection of compounds of interest. As proof of principle we correctly classified all subjects from population of smoking (N=11) and non-smoking (N=11) based on the VOC profiles available in their exhaled air. Support vector machine (SVM) analysis identified 4 VOCs as biomarkers of recent exposure to cigarette smoke: 2,5-dimethyl hexane, dodecane, 2,5-dimethylfuran and 2-methylfuran. This approach contributes to future development of fast, accurate and non-invasive diagnostics of inflammatory diseases including pulmonary diseases.