Abstract: Stable isotope analysis has emerged as one of the primaries means for examining the structure and dynamics of food webs, and usually used to estimate the contribution of various sources to consumer diet. The existence of outliers in the data, the lack of prior information, and the unreasonable interpretation of the posterior distribution could make an inaccurate model to explain the contribution of consumer diet. Where possible, the application of stable isotope data should be supplemented with additional data such as diet analysis and feeding behavior. To this day, there are still many methods that require specific information about the consumer diet, which is used as model parameters and the establishment of prior information in the Bayesian mixing model to reduce the number of potential source pools. Current models estimate probability distributions of source contributions and incorporate complexities such as variability in isotope signatures, discrimination factors, hierarchical variance structure, covariates, and concentration dependence. It is important to incorporate sources of variability and uncertainty and not just rely on point estimates in mixing model inputs. Outputs from Bayesian models are true probability distributions that may be plotted and summarized with any number of descriptive statistics, as well as compared with each other, with hypothesized distributions, or with other parameters of interest such as fitness. Regardless of the method of analysis, stable isotope data can only indirectly reflect the flow of energy and nutrients in the food web, and can not directly provide clear information about the functional relationship between organisms. Standardized and simple statistical testing of original data provide a reasonable way to evaluate the uncertainty of instrumental analysis, and improve data quality and eliminate abnormal data; analysis of isotope space helps to verify whether the data quality conforms to the modeling need; Prior information can accurately reflect the feeding relationship combined with reasonable interpretation of the posterior distribution, which can not only improve the accuracy of the mixed model prediction, but also trace the source of consumers nutrition to the maximum. The acquisition of empirical data such as nutritional identification factors, dietary prior information, tissue turnover rate, etc, still requires a lot of experimental support in the field and lab. In addition, according to specific research, in the course of practice, there are various evaluation indicators, and they respectively describe the information loss relative to the “real model”. Due to the unknown nature of the real model, these evaluations can only reflect the relatively good performance of the existing model construction process, and specific issues still need to be analyzed in detail. Therefore, this paper created a series of Bayesian models based on the measured isotope data set (isotope data set for plankton), the simple statistical test of the original data and the correction of the prior information of nutritional sources, the construction and test of the isotope space, as well as the prior information, various analysis methods and steps, such as the difference in posterior distribution, to describe the method and process of consumer nutrition source traceability, which provide guidance methods for the application of stable isotope technology to carry out consumer nutrition traceability research.