Abstract: Large-scale pressure vessels are widely used in petrochemical, thermal power, nuclear energy, and other fields. As they are usually used in high-pressure, high-temperature, and corrosive environments, periodic testing is a necessary measure to ensure engineering safety. However, the internal and external environment of pressure vessels is complex and difficult to detect. Traditional testing methods often cannot meet the needs of efficient and accurate detection. To this end, robotic systems developed based on technologies such as robotics, computer vision, and artificial intelligence have become an effective approach of nondestructive testing. This paper focused on the development status of weld nondestructive testing robot in special operating environments such as work at heights and confined spaces, introduced the movement methods, defect data acquisition and identification technologies then discussed the application status and development prospects of artificial intelligence in defect identification.