Aiming at the user privacy protection technology in released digital library data, we propose a personalized anonymous method. The user sets sensitive factors by themselves, and assigns attribute weights according to association rules among data attributes, thus obtaining the user information privacy protection degree of the data. We utilize this degree to conduct data division and anonymity, and realize personalized anonymous protection for the user. Experimental results show that the personalized parameters obtained through the weights of attributes can better reflect actual data relationship, decrease "excessive" protection caused by users' personalized settings, and in the meantime improve data publishing quality.