In recent years, cross-site scripting attacking usually occurred on the social networks.The attackers hide the malicious script codes in web links, and easily get personal sensitive information if someone clicks these links.This article firstly analyzes the characteristics of the social network topology and the relationships of network users, and then establishes the worm propagation model in order to draw out worm prop agation procedures among network nodes.Finally some impacts, which consist of the number of online users, theprobability of users click worm links, the release time of immune patches and the quantity of users installed patches on the worm prop agation, are discussed.