As one of the most critical components in rotating machinery, bearings play a vital role in ensuring the stable operation of the equipment. Due to the influence of complex environments, there is uncertainty in the process of bearing fault diagnosis. The Belief Rule Base (BRB) is a modeling method that can effectively handle uncertainty issues. However, in the process of bearing fault diagnosis, with the increase of bearing use time, bearing wear will become more and more serious. When extracting data feature values, different types of fault feature values have similarities, which makes it difficult to distinguish and represent the type of fault in the diagnosis process, resulting in local ignorance. The traditional BRB model cannot solve the local ignorance problem, and with the increase of the number of rules, it will lead to the explosion of rule combinations. To solve the above problems, this paper proposes a Hierarchical Belief Rule Base with power set, Firstly, a Hierarchical Belief Rule Base (HBRB) model is selected to solve the problem of combination rule explosion, and then HBRB is extended to the power set identification framework. Because the power set recognition framework can more effectively represent the ignorant information in the complex system, and solve the local ignorance problem in the bearing fault diagnosis, then, the evidential reasoning algorithm is used to establish the reasoning process of the model. Finally, the parameters of the model are optimized using the projection covariance matrix adaptive evolutionary strategy algorithm. The experimental results show that the proposed method can identify different fault categories with high accuracy, effectively solve the local ignorance problem in bearing faults diagnosis, and has good generalization ability.