In view of the rapid decline in nutrition during transportation of crabs and the serious loss of quality, the problem of effective monitoring is difficult. The biosignal detection device was used to obtain the changes of gas and bioimpedance of crabs at 4℃, 15℃ and 25℃. The chemical experiment was used to detect the changes of volatile base nitrogen in crabs in the same environment. According to the change of sensory acquisition during crab storage, the quality change stage was roughly divided into three stages: fresh to sudden death (S1), sudden death to death (S2) and postmortem corruption (S3). The BP neural network was optimized by genetic algorithm (GA), and the coupling model of impedancevolatile base nitrogen, gasvolatile base nitrogen and impedancegasvolatile base nitrogen was established. The results showed that the prediction accuracy of the model at 4℃ was 96.93%, 86.32% and 98.77%, respectively. The prediction accuracy of S1 at 15℃ was 8632%, 86.32% and 87.04%; S2 was 87.04%, 88.86% and 89.81%; S3 was at least 92.28%, 96.20% and 93.79%. The prediction accuracy of S1 at 25℃ was 88.86%, 87.04% and 91.23%; S2 was 86.32%, 87.72% and 93.79%; S3 was at least 89.81%, 89.35% and 86.32%, respectively. The coupling model basically met the needs of quality prediction of river crab. The quality of river crab can be predicted through the biological signals which was done by building the link between volatile compounds, biological impedance and chemical indexes.