Since the instability of dangerous rock mass exhibits as the sudden collapse of destruction, no obvious displacement characteristics can be identified. Thus, the application of conventional displacement monitoring techniques is difficult to achieve the purpose of monitoring and early warning. In addition, the stability of dangerous rock mass is influenced by the bonded area between dangerous rock block and bedrock, which is considered as one of the critical parameters. However, this bonded area is hard to obtain, which makes it difficult to evaluate the stability of the dangerous rock block. In this study, we assume that the rock block is homogeneous and isotropic, the main control structural plane is a single plane, the damping ratio of the system is less than 1 and the deformation is linear elastic deformation within the amplitude range. As a result, the vibration model of the dangerous rock mass can be simplified as a spring oscillator model. Then, the relationships among the natural vibration frequency of dangerous rock mass, the bonding area, the elastic modulus and the quality of dangerous rock mass were established by the theoretical derivation. Considering the limit equilibrium model, a new model based on natural vibration frequency was derived to evaluate the stability of dangerous rock mass. A dangerous rock block on the right bank of Baihebao reservoir was selected as the case study. A wireless vibration sensor (micro core) was fixed on the dangerous rock block to acquire data, which was converted to natural vibration frequency by the Fourier transform and other mathematical methods. Furthermore, the bonded area was determined according to the relationship among natural vibration frequency, rock block bonding area and elastic modulus. Finally, the stability evaluation of dangerous rock mass was completed using the new model. This case study verifies the feasibility of the proposed procedure with a faster rate and more accuracy than traditional methods.