With the rapid development of mobile wireless technology, mobile social networks play a key role in people’s online life. However, a large amount of data containing indi-vidual relationship information in mobile social networks will result in the leakage of in-dividual privacy. Therefore, how to prevent privacy disclosure of these network data while sharing them to improve services for users is urgent. In order to improve the effec-tiveness of differential privacy, an uncertain graph method based on the shuffle model is proposed. Especially, the shuffle model is introduced to modify the relationships of nodes, which not only provides differential privacy preserving for the link privacy of nodes, but also improves the data utility of differential privacy. Moreover, node differen-tial privacy is utilized to inject uncertainty on edges, which can reduce perturbations caused by differential privacy. In addition, the exponential mechanism is used to restrict the edge modification in the original graph. The theoretical analysis shows that the un-certain method satisfies differential privacy. The results of experiments show that the uncertain method can effectively preserve link privacy of nodes and maintain data utility.