Abstract: In order to effectively control, manage and reduce the adverse effects of highway slope geological disasters, this study proposes a new method of highway slope risk assessment based on UAV tilt photography and three-dimensional real scene slope information extraction combined with analytic hierarchy process, and develops a three-dimensional visualization web platform for highway slope risk assessment along the Yellow River Based on cesium. The three-dimensional real scene model and computer algorithm can be used to extract the slope risk evaluation information (including road slope distance, slope, slope height and joint density, etc.) in the room. The AHP is used to evaluate the risk of each slope. According to the comprehensive final risk score, it is divided into five risk levels: extremely low, low, medium, high and extremely high. The proposed method is applied to the risk assessment of slopes along the Yellow River Highway. A total of 656 slopes are evaluated, including 0 extremely low-risk slope, 23 low-risk slope, 405 medium risk slope, 210 high-risk slope and 18 extremely high-risk slope. The evaluation results are consistent with the field survey results, which show that the evaluation system proposed in this paper is effective and reasonable. The system not only realizes the three-dimensional visualization investigation and risk assessment of geological disasters, but also promotes the development of digital and intelligent application of highway slope disaster prevention and reduction technology.