Colorimetric sensor array system was employed to characterize changes in volatile organic compounds (VOC) in aquatic products (prawn, small yellow croaker, shuttle crab) with different storage time. Firstly, the VOC of different aquatic products were characterized by colorimetric sensor array system, and the trends of VOC changes during storage were analyzed by principal component analysis (PCA). Linear discriminant analysis (LDA) qualitatively identified the freshness of prawn, shuttle crab, and small yellow croaker. The results showed that the recognition rates of fresh prawn, stale prawn, fresh small yellow croaker, stale small yellow croaker, fresh shuttle crab, and stale shuttle crab were 94.44%, 93.75%, 95%, 100%, 100%, and 92.31%, respectively. The technique was then combined with the back propagation artificial neural network (BP-ANN) model to quantitatively predict the amount of TVB-N in these products. It was found that the correlation coefficient values between this model and semimicro-kjeldahl determination to determine TVB-N levels in prawn, shuttle crab, and small yellow croaker were 0.988 4, 0.995 4, and 0.983 8, respectively. These results showed that colorimetric sensor array can be used to characterize the freshness of aquatic products rapidly.