Select Shewanella putrefaciens isolated from the fish as the research object， explored the effect of environmental factors including pH， aw and NaCl on its growth probability at room temperature （25 ℃）. The second order linear logistic regression equation and the PNN artificial neural network algorithm were used to establish the growth / non - growth interface model of Shewanella putrefaciens， the goodness of fit and the predictive power of the two models were compared and validated. The results showed that the coincidence rates were 93.80% and 100.00% of the trainingset and test set of the second order linear logistic regression equation which could predict the growth probability in detail； while the consistent rate of trainingset and test set of PNN artificial neural network is 100.00% and 86.67% and the optimal speed is 0.1 second， which can classify the growth / non-growth data. With the increasing of salt， the growth of Shewanella putrefaciens was inhibited， the growth interface moved to high water activity， high pH direction； The spoilage organism couldn’t grow under the condition of water activity of 0.91， aw=0.92， 0.94 and 0.96， growth probability of Shewanella putrefaciens increased with the increase of pH， rise steeply； while pH=4.5， Shewanella putrefaciens did not growth， with the increase of pH， the growth probability of high aw cases first reached 1. By developing the Shewanella putrefaciens growth / no growth interface model under different environment parameters， PNN artificial neural network model could make a good andrapid classification predictionon the growth / no growth data； The second order linear logistic model was better for the quantitative assessment of pH， aw and salt domains to protect the quality and the stability of products.