The invention discloses a risk management and control system based on scene recognition driving. The risk management and control system comprises a data resource library, a patrol terminal system, a risk management system and an interface standard specification. The safety production risk grading management and control system is based on a data resource platform. Based on SOA-based architecture design, frontier technologies such as wireless communication, geographic information service, big data analysis and computer network security are fused, system data transmission is achieved through local area network (enterprise intranet) and 4G/WIFI network communication, two functional interfaces of mobile terminal application and platform service application are formed based on application support, a standard multi-level system is adopted, interfaces are strictly defined among all levels, and system seamless integration is achieved through overall design and deployment. The system can be applied to multi-platform equipment such as a mobile phone, a tablet personal computer and a computer; and the system can be applied to high-risk environments or places, can be used remotely through a public wide area network, can also be used through a local area network, collects dangerous point data through mobile inspection equipment, and uploads the dangerous point data to a database in time, sothat a data resource library can update dangerous information data in time, and an enterprise is helped to dynamically master safety production dynamic information anytime and anywhere. Meanwhile, therisk management system carries out early warning management according to the data of the data resource library, and predicts the risk accident trend by adopting artificial intelligence in combinationwith a data mining analysis method, thereby helping enterprises to make effective decisions. The system can also be applied to the field of fire protection, and help emergency departments to manage and control risks and respond timely.