Landslide damming occurs frequently, affects a wide range of areas, and is difficult to predict, thus threatening the safety of human life, properties, and engineering structures. However, the studies on early identification remain at an early stage. This paper proposes a hazard prediction method for landslide damming, which includes prediction models of river blockage, landslide dam geometry, landslide dam stability, and dam breach time. On this basis, the landslide damming hazard can be predicted via ten parameters: the repose angle of the landslide mass, the landslide volume, the landslide discharge, the river discharge, the reservoir capacity of the barrier lake, the bottom width of the valley, two slope angles on both sides of the valley, the valley bed inclination, and the sliding surface dip-angle. The proposed method was examined in relation to a typical case, namely, the Baige landslide damming that occurred in 2018 in China, with the application process presented integrally. The results indicated that the river blockage model predicted that the Baige landslide could form a huge landslide dam, which was fully in line with the actual situation. The landslide dam geometry model provided the predicted value of the height, width, and length, as well as the downstream and upstream angles, of the Baige landslide dam, with the maximum relative error between the predicted values and the true values less than 20%. The prediction results of the landslide dam stability model indicated that when the reservoir capacity of the barrier lake is 280 million m3, the dam could breach. In fact, the dam breach time model predicted that the dam could breach within 38 h after the landslide dam formation, which was 11% less than the actual breach time of the Baige landslide dam. The current study provides a useful theoretical reference and technical support for the early identification, prediction, and warning of the disaster chain of landslide-induced river blocking.