Portable electronic nose (e-nose) consisting of polymer/carbon nanotube (CNT) sensor array was developed to detect protein-based foods. Gas sensors were fabricated by spin coating functionalized CNT/polymer nanocomposite materials onto interdigitated electrodes. The sensors were tested with various types of volatile compounds such as ammonia, amine compounds, acetic acid, water and organic solvents in the range of ppm level. It was found that most sensors yield strong signals to ammonia, amine compounds and acetic acid as well, while they present quite low response to organic solvents and water. To understand the relation of the interaction of amine species related to the sensor response, we have performed molecular modelling based on the density functional theory (DFT) on one polymer structure providing the best response to volatile ammonia. Based on the principal component analysis (PCA), this portable e-nose was successfully applied for the classification of the seafood releasing different amount of amine compounds.