Abstract: To improve the trust evaluation accuracy of the DyTrust model, solve the problem of rough granularity of the trust model, and deal with the trust evaluation problems caused by individual experience, based on the DyTrust model, an improved algorithm of trust evaluation was presented after using an experiential factor and the particular services of nodes. Compared with the DyTrust model, the improved model has the advantages of crisp granularity, high accuracy, reflecting the personalization of nodes, enhancing the feedback trust value, and high scalability.