Abstract: With the in-depth deployment of China’s “Belt and Road Initiative” and other major strategies, major national strategic projects, such as the Sichuan–Tibet Railway, will be built in mountain valleys with complex terrain. Under the influence of continuous external factors (rainfall, earthquake, construction, etc.), rock masses of a high-steep rock slope in a stable state gradually transform to unstable rock, and then the unstable rock slides and destroys, which is a major hidden danger threatening the safety of project construction and operation. Unfortunately, the stability factor (SF) can only identify the occurrence of failure and cannot identify the transition of stable to unstable rock; thus, it cannot quantitatively identify unstable sliding rocks. Rock mass failures mostly evolve from unstable rocks. Therefore, how to establish a quantitative evaluation method to identify unstable sliding rock masses is a major problem in the field of early warning and prevention of rock collapse. In this study, the potential sliding surface cohesion and its slip resistance share are analyzed by introducing a cohesive stability factor (CSF) to achieve dynamic consideration of the stability, separation, and damage phases of slip-type rock masses. When the CSF is <1 and the ratio of cohesion to skid resistance (η) drops below the ratio of long-term strength to failure strength, the sliding rock masses become unstable. Among them, the CSF is the main control indicator, and η is the auxiliary indicator. The laboratory experimental results of sliding unstable rock mass show that a single SF cannot accurately identify the A8 unstable rock mass in the experimental group quantitatively because of the inconsistency of the standard, and a quantitative identification method of unstable rock masses based on the CSF and SF can scientifically realize the quantitative evaluation of unstable rock masses in terms of mechanics. The field case study shows that the improved method provides relatively objective judgment criteria for the quantitative identification of unstable sliding rock masses with an SF of ~1.2 in the Chongqing Three Gorges and other regions. Compared with the traditional mechanical identification methods of unstable rock masses, the new identification method proposed in this study can conduct a set of objective and unified identification criteria, which improves the accuracy and scientificalness of the traditional mechanical identification methods and provides an effective reference for better management of rockslides in high-risk areas.