Recommender systems are used with the purpose of suggesting contents and resources to the users in a social network. These systems use ranks or tags each user assign to different resources to predict or make suggestions to users. Lately, social tagging systems, in which users can insert new contents, tag, organize, share, and search for contents are becoming more popular. These systems have a lot of valuable information, but data growth is one of its biggest challenges and this has led to the need for recommender systems that will predict what each user may like or need. One approach to the design of these systems which uses social environment of users is known as collaborative filtering (CF). One of the problems in CF systems is trustworthy of users and their tags. In this work, we consider a trust metric (which is concluded from users tagging behavior) beside the similarities to give suggestions and examine its effect on results. On the other hand, a decentralized approach is introduced which calculates similarity and trust relationships between users in a distributed manner. This causes the capability of implementing the proposed approach among all types of users with respect to different types of items, which are accessed by unique id across heterogeneous networks and environments. Finally, we show that the proposed model for calculating similarities between users reduces the size of the user-item matrix and considering trust in collaborative systems can lead to a better performance in generating suggestions.