Abstract: To investigate the causes of fish spoilage under different temperatures and develop the predictive model of shelf life, sensory, chemical and microbiological characteristics and bacterial flora of cultured Pseudosciaena crocea stored aerobically at low temperature from 0 to 10℃, ambient temperature at 25℃ and non-isothermal temperatures were carried out, and the specific spoilage organisms in cultured Pseudosciaena crocea at above mentioned temperatures were also identified. And then three predictive models of shelf life were developed and validated under isothermal and non-isothermal conditions. The results showed that the shelf lives were from 5.4 to 17.8 days and their specific spoilage organisms were Shewanella putrefaciens, Pseudomonas at low temperature, and shelf life was only 1.1 days and specific spoilage organisms were Vibronaceae and Enterobacteriaceae at ambient temperature. Predictive models of shelf lives including Exponential, School-field and Square-root equations were fitted based on the relationship between relative spoilage rates and temperatures, and three model parameters which were the temperature characteristics coefficients(a), minimum temperature(Tmin), apparent activation energy (Ea) were 0.11, -10℃and 74 kJ/mol, respectively. Three predictive models of shelf life were validated under isothermal (3,7and 10℃) and non-isothermal conditions. According to the comparison of relative errors of three models which were 0%-13.8%, -0.9%-9.8% and -0.2%--22.1%, it showed that School-field and Exponential models of shelf life were better than Square-root model for evaluating quality of fish at 0-25℃. This paper provides a basis for developing the intelligent predictive systems of quality integrated with the package, storage and distribution of fish products.