In this paper, a portable electronic nose (Enose) system was designed, and its application in K value prediction for large yellow croaker (Pseudosciaena crocea) was also explored. E-nose responses to croaker samples were measured for 8 days. Microbiological and chemical indexes including total viable count (TVC), K value, and GC-MS were synchronously examined. Principal component analysis (PCA) and stochastic resonance (SR) were utilized for E-nose measurement data analysis. Results suggested that croaker's K value increased rapidly due to microbial propagation. GC-MS results presented the information of volatile gas emitted by croaker samples, and provided reliable references for E-nose responses. PCA method could not discriminate all croaker samples, while SR signal-to-noise ratio (SNR) maximum (Max SNR ) values quantitatively discriminated all samples. Large yellow croaker K value predictive model was developed by linear fitting regression between K values and Max SNR values with a regression coefficient (R 2 ) of 0.96. Validating experiment results demonstrated that the forecasting accuracy of this model is 83 % the regression coefficient of the developed model was 0.83. E-nose is a useful tool for food quality examination. But, there are no such standard references for E-nose examination, which limits the applications of this technique. The method proposed here will promote the applications of E-nose in aquatic product quality rapid analysis.