The assessment of microbial spoilage in fresh fish is a major concern for the fish industry. This study aimed to evaluate the efficiency and reliability of an electronic nose (E-nose) to detect microbial spoilage of fresh sardines (Sardinella longiceps) by comparing its measurements with Total Bacterial Count (TBC), Hydrogen Sulfide (H₂S) producing bacterial count and Trimethylamine Oxide (TMAO) reducing bacterial count after variable storage conditions. The samples were stored at 0 °C (0, 2, 4, 6, and 8 days) and 25 °C (0, 3, 6, and 9 h), while day 0 was used as a control. The E-nose measurements were analyzed by Principal Component Analysis (PCA), Linear Discriminant Analysis (LDA) and Artificial Neural Network (ANN). Microbial counts increased significantly and simultaneously with the changes in E-nose measurements during storage. The LDA and ANN showed a good classification of E-nose data for different storage times at two storage temperatures (0 °C and 25 °C) compared to PCA. It is expected as PCA is based on linear relationships between the factors, while ANN is based on non-linear relationships. Correlation coefficients between E-nose and TBC, TMAO-reducing bacterial and H₂S-producing bacterial counts at 0 °C were 0.919, 0.960 and 0.915, respectively, whereas at 25 °C, the correlation coefficients were 0.859, 0.945 and 0.849, respectively. These positive correlations qualify the E-nose as an efficient and reliable device for detecting microbial spoilage of fish during storage.