Abstract: Existing anonymous models can overgeneralize sensitive value data, which will result in more loss of information and insufficient consideration for personalization of sensitive attributes. In response to the above questions, (λα,p,k^m) hierarchical personalized anonymous model is proposed, and an anonymous algorithm CPPAA is designed to satisfy the model. Considering the degree of privacy protection of sensitive attributes, the model ranked the sensitivity of sensitive attributes and required the number of categories of sensitive attribute values. It considered the anonymous processing of transaction operations in some data sets, which reduced the risk of privacy disclosure to a certain extent. In addition, the particle swarm optimization (PSO) was introduced to further optimize the search process for optimal records. The experimental results show that this method can protect sensitive attributes and reduce the loss of information as much as possible.