Penetration testing is a method of assessing the security of a system, application, or network by simulating the actions of an attacker. With the continuous development of the digital economy, the data protection and offensive and defensive confrontation of digital economic infrastructure are becoming increasingly fierce. How to improve the efficiency and accuracy of penetration testing through automation has become increasingly important. As an important part of artificial intelligence technology, knowledge graph provides new ideas for automated penetration testing with its efficient semantic understanding and reasoning capabilities. This paper can be divided into three parts according to the main content. First, it reviews and analyzes the current research status of automated penetration testing path planning. Secondly, it introduces the related construction methods of cyber security knowledge graph and ontology. Finally, it analyzed and compared attack path generation technology from multiple perspectives and looked forward to future research directions.