Abstract: The unimpeded operation of the drainage channel is critical to ensuring the flood season safety and overall stability of tailings ponds. Traditional manual inspection methods, characterized by low efficiency and incomplete coverage, are inadequate to meet the demands of tailings pond safety management. To address this practical challenge, this paper proposes an intelligent identification method for drainage channel blockages by integrating UAV photogrammetry technology with the YOLO deep learning model, aiming to achieve autonomous detection and information acquisition of blocked drainage channel sections. High-resolution digital orthophoto maps (DOM) of the pond area are acquired via UAV, cropped into several 30 m × 30 m standard grid units, and numbered accordingly. A YOLO11 m-OBB rotating target detection model is constructed to iteratively detect blockages in each grid unit, automatically annotate the blocked sections of drainage channels, and extract blockage information in conjunction with the DOM imagery. Validation was conducted using two tailings ponds in Northeast and North China as case studies. The results demonstrate that the method successfully identified 9 and 12 instances of drainage channel blockages, respectively. By integrating image geographic information, the projected coordinates and lengths of the blocked sections were calculated, achieving comprehensive and rapid inspection of drainage channel blockages. The findings provide robust information for precision on-site clearance operations, enhancing the timeliness and targeting of drainage channel maintenance. This research offers an intelligent and efficient technical means for the daily management of tailings ponds and the investigation of hidden disaster-causing factors. It holds significant engineering application value and promotion prospects for ensuring the long-term stable operation of the drainage channel and securing the flood control safety of tailings ponds.