In order to study the changes of taste substances of Fugu bimaculatus during cold storage, a rapid method to detect the freshness of Fugu bimaculatus during cold storage (0 ℃ and 4 ℃) was established. The results of total bacterial count (TVC) were used as the criterion of fish freshness, and the response value of electronic tongue sensor was used as the dependent variable. The partial least squares (PLSR) and multiple linear regression (MLR) methods were used to establish the TVC prediction model, which can be used to distinguish the freshness of Fugu bimaculatus in different cold storage periods, and the model was verified. The results showed that both PLSR and MLR models can predict the freshness of fish at different refrigeration periods, but the MLR model had a better fit. The training sets (Rc2) of the TVC prediction models for the 0 ℃ and 4 ℃ groups were 0.98 and 0.99, respectively, The coefficient of determination (Rv2) of the prediction set was 0.97 and 0.99, the root means square error (RMSEV) of the training set was 0.40 and 0.08, and the root mean square error (RMSEP) of the prediction is 0.44 and 0.08. External verification results showed that the accuracy of the model test set was 100%, and it had a good predictive ability. The above results showed that the electronic tongue technology could be used as a rapid detection method for the freshness of the Fugu bimaculatus in the cold storage process. Taken together, these results suggest that electronic tongue technology can be used as a rapid detection method for the freshness of the Fugu bimaculatus in the cold storage process.