For the large group decision making problem with incomplete decision preference information, the trust mechanism belonging to the access control is introduced, the direct trust degree and recommendation trust degree are established, and a compensation method based on the trust mechanism is proposed. With analyzing the problems of previous distance similarity, a new distance formula of similarity is proposed, which is combined with the cosine similarity, and a 2-tuple similarity model to clustering is established. Based on the proposed clustering model, the decision makers’weight is solved by using the clustering method, and the decision ranking is obtained by the synthesis of weights and the complete preference matrix. Finally, an existing literature case is given to illustrate the effectiveness and advantage of the proposed method.