The invention discloses a mixed randomization privacy protection method of social network data dissemination. The mixed randomization privacy protection method includes the steps that top points V of an undirected graph G(V,E) for expressing social network data are divided into a set Vk and a set Vr according to degree numbers with k-anonymity as a condition, wherein the set Vk comprises the points with the degree numbers meeting the k-anonymity condition, and the set Vr comprises the points with the degree numbers not meeting the k-anonymity condition; the points in the set Vk are not processed; randomization protection processing is carried out on the points in the set Vr. According to the mixed randomization privacy protection method, an existing k-anonymity method and an existing randomization method are combined, the number of users of the social network is not changed, influences on the nature and the quality of data provided for third parties are small, the processing process is simple, and the processing speed is high; after processing is carried out, the social network data of each person can be in one of the two following protection states: the degree natures are not changed, and the social network data of the person are located in other at least k persons same as the person in degree nature, or the degree natures are changed; relocation attacks with structure background knowledge can be prevented through the protection, and the demands of the users of the social network for the privacy safety are met.