Conflict analysis has become a hot issue in management science. In the context of conflict analysis, there are three attitudes for agents to describe the opinion, including supportive, opposite, and neutral. Then, the conflict situation is discussed and analyzed. In this paper, we propose an extended Pawlak conflict model concerning the trust mechanism to solve the problem of the reaching consensus process. Firstly, the degree of conflict is defined by the weights of agents considering the penalty factors, and an extended Pawlak conflict model is presented. Then, the trust recommendation mechanism is proposed to modify the opinions of agents and reach conflict consensus. Four kinds of feedback mechanism are discussed by using four perspectives: 1) without the penalty factors and no limit to the range of adjustments; 2) without the penalty factors and the attitude of agents vary from pessimistic to neutral; 3) with the penalty factors and no limit to the range of adjustments; 4) with the penalty factors and the attitude of agents vary from pessimistic to neutral. Furthermore, this paper presents a process of reaching consensus based on the trust recommendation mechanism for the conflict analysis problem. Finally, a case study is used to validate the effectiveness and superiority of the proposed method. A comparative analysis is completed for the parameters and the maximum alliances could be obtained with the original Pawlak conflict model. (C) 2021 Elsevier Inc. All rights reserved.