Abstract: In geotechnical engineering practice, an important task is to obtain the design parameters required based on the site survey report. The mutual conversion of soil parameters is of great significance to the practice of geotechnical engineering. Based on the previous studies, the laboratory test and field test data of soft soil in a certain urban area of Zhuhai are analyzed. A multivariate probability distribution model of 9 soil parameters and a Bayesian updating model are established. Using this Bayesian updating model can easily make use of the information of multiple soil parameters to predict unknown parameters. The results show that the multi-parameter prediction model can synthesize the information of all known parameters, thereby effectively improving the accuracy of predicting unknown parameters.