<p>Maintenance has always been an important part of a mechanical equipment system. Inspection, as part of the activities of maintenance, is conducted by human inspectors whose dependability is fundamental to an effective maintenance system. The inspector should have the superior decision-making ability to deal with unforeseen events in an uncertain environment. This paper proposed a multi-scenario simulation method for authoring an immersive virtual reality (VR) system for inspection training of hot strip rolling mills in order to cost-effectively simulate the uncertainty in the types and locations of equipment defects. To begin, an efficient representation of defect characteristics is developed to model existing and emerging defects. Concurrently, the defect model files and properties are saved and managed in the defect model library to be reused to construct virtual environments (VEs) for various inspection scenarios. Second, the corresponding defect equipment under inspection is established based on the information of inspection scenarios by modifying the original equipment assembly structure, which is built in advance with domain knowledge. The VR system can auto-build VE based on the scene and assembly structure information configured by the training personnel and defect library. Finally, the multi-scenario inspection training system of hot strip rolling mills was applied to the 9th national iron and steel industry vocational skills competition. Following the competition, more than ten companies have implemented the system to train technicians, which improves the training experience and effectiveness.</p>