User-generated content has become more and more popular. The success of collaborative content creation such as Wikipedia shows the level of user's accomplishments in knowledge sharing and socialization. In this paper we extend this research in the service domain, to explore users' social behavior in Collaborative User-Generated Services (Co-UGS). We create a model which is derived from a real social network with its behavior being similar to that of Co-UGS. The centrality approach of social network analysis is used to analyze Co-UGS simulation on this model. Three Co-UGS network actors are identified to distinguish users according to their reactions to a service, i.e. ignoring users, sharing users and co-creating users. Moreover, six hypotheses are proposed to keep the Co-UGS simulation. The results show that the Co-UGS network constructed by the sharing and co-creating users is a connected group superimposed on the basis of the social network of users. In addition, the feasibility of this simulation method is demonstrated along with the validity of applying social network analysis to the study of users' social behavior in Co-UGS.