Abstract: In order to strengthen the monitoring on water conservancy projects during engineering construction and operation period, master the engineering construction progress in real time, and realize the normalization, automation and intelligent inspection of water conservancy projects, based on the advanced technologies and methods such as UAV, Internet of Things, 4G/5G, GIS, deep learning and target detection algorithm, we study the integration technology of UAV automatic inspection system, multi-channel remote video transmission technology, image abnormal feature object recognition technology and orthophoto classification and recognition technology. Then the automatic inspection technology system with UAV was constructed, and the intelligent monitoring system with UAV automatic inspection was developed, which is applied to the intelligent inspection in Datengxia Water Control Project. The practical application results showed that this system realized more than 90% of the illegal sensitive point recognition and the storage management of massive inspection big data in batches. The inspection efficiency was increased by more than three times, and the success rate of image recognition was increased from 59.83 % to 90.17 %, also the efficiency and success rate of building recognition of orthophoto were greatly improved. This method can provide real-time visual inspection data for the dynamic supervision of water conservancy projects during the construction period, and also provide strong technical support for the supervision works in water conservancy industries such as flood control emergency rescue and law enforcement in river shoreline.