In order to solve the problem that the classical Dempster-Shafer (D-S) evidence theory combination formula fails when there is a high conflict between evidence and increase the convergence of conflicting evidence fusion, the unified trust distribution mechanism and the reward-punishment mechanism are proposed from the perspective of mutual trust between basic probability assignments (BPAs). The unified trust distribution formula is proposed as a framework for calculating the weights of BPAs. The reward-punishment factors are proposed to differentiate the importance of different BPAs. The unified trust distribution formula and the reward-punishment factors are combined to correct and fuse the evidence. Some numerical examples and two practical application examples are used to illustrate the applicability and superiority of the proposed method. In the fusion optimization of abrasive media, the BPA of the abrasive media actually used is 0.839. In an application for target recognition, the BPA of the actual target is 0.9971. The experimental results validate that the proposed method is more effective compared to other methods. The proposed method focuses on identifying and correcting evidence from the level of BPAs, and improves the convergence of fusion results, which provides a new research idea for improving the evidence combination method.