Abstract: Taking 2 249 shipping service enterprises in 13 cities of Jiangsu Province as the research object, the index system covering port operation, economic development and social environment is constructed. In order to eliminate multicollinearity and ensure the fitting effect, Bayesian ridge regression model is used to study the mechanism of agglomeration of different types of shipping service industries. The results show that the clustering effect is obvious in coastal and riverine areas; low-end and middle-end shipping service enterprises have similar spatial clustering power mechanism and are most significantly affected by the port operation; high-end shipping service enterprises are more sensitive to the social environment, economic development and other indexes, and other port operation indexes other than container throughput do not have a significant impact on them; localities should find the focus of the development of the shipping service industry in accordance with the local conditions and formulate differential support policies to promote the upgrading of the port and shipping industry.