In view of the engineering problem of port construction in the area with poor geological condition,poor or even bad sea condition,the object-oriented method is adopted to extract the knowledge of construction entities and establish the construction knowledge graph system of pile-supported wharf.The key information of construction knowledge of pile-supported wharf is represented by triplet,and the knowledge fusion is carried out according to three levels of knowledge graphs of prefabricated component,construction process and quality control requirements by the entity alignment method of graph convolutional network.The construction factor similarity calculation method is used to calculate the Levenshtein distance between the target data and the typical process condition data in the knowledge graph under specific engineering conditions,and the overall construction process flow is determined.The constraint rule decision tree is used to impose constraints on the construction body in terms of production resources,technical standards,safety risks and comprehensive evaluation,and key construction parameters in the specific process are clarified,then the logical route of construction technology is optimized.The Unreal Engine is used to create the offshore construction inference system of pile-supported wharf,and the intelligent planning of pile driving process and the intelligent decision of pile stopping condition under different construction conditions are realized by designing blueprint script.