While the rapid increase in unmanned aerial vehicles (UAVs) prompts the development of the low-altitude economy, the possibility of conflicts between UAVs also increases, which poses threats to operation safety. Therefore, it has become an urgent need to quickly and accurately detect conflicts between UAVs. This paper proposes a visual detection method for UAV conflicts based on YOLOv5 and monocular ranging approach. A lightweight YOLOv5s model is used to detect UAV targets that are then processed by a monocular ranging algo- rithm to measure the relative distance between UAV. The UAV conflict is iden- tified by comparing the distance threshold and the measured distance between UAVs. The experimental results show that the proposed model reaches a preci- sion of 95.82%, a recall rate of 95.78%, and an average precision of 95.64%. This demonstrates that the UAV conflicts can be effectively detected by the proposed model.