In the contemporary intelligent and digital landscape, unmanned aerial vehicles (UAVs) have become integral to daily operations, engaging in urban surveillance, weather forecasting, and agricultural monitoring. Nevertheless, due to UAVs’ diminutive size, diverse types, and the intricate conditions of urban environments, achieving real-time detection of UAVs within cityscapes presents substantial challenges. Furthermore, as the number of UAVs proliferates, devising air trafffc management systems to ensure urban safety and enhance the urban environment becomes paramount. Consequently, the development of a novel system tailored to the complexities of urban environments, capable of accurate UAV recognition and continuous tracking, is imperative. In this research, we employ the cutting-edge YOLOv5 as the target detector, coupled with the DeepSORT target tracker, to introduce an innovative UAV auto-recognition and tracking scheme. This scheme not only facilitates the precise identiffcation of small UAVs amidst complex urban settings but also efffciently differentiates various UAVs while maintaining consistent tracking. The results demonstrate that this approach signiffcantly augments the stability and reliability of the tracking process, offering substantial practical value.