Abstract. The existing research of churn prediction in social network mainly consider single individual behavior and adjacent relationship structure, we propose a churn prediction iterative model in social network based on relationship strength. First of all, according to the traditional model, we put forward a hierarchical structure approach to measure the strength of social relationship based on the traditional user relationships and the complexity of social theory. Secondly, from the two angles of the user's activity and user's influence, we apply iterative way to calculate the influence factor of the social environment. And based on conditional probability model, we put forward a churn prediction iterative model in social network (CIM). Experiments show that the model is suitable for the churn prediction in social network with a good effect, and improve the accuracy on churn prediction in social network.