Abstract: The operation health of transmission pipeline and equipment affects the safety production of coal-fired power plant. At present, there are still some problems in the wear prediction and maintenance management technology of power plant pipeline and equipment at home and abroad, which are mainly characterized by small fault diagnosis range, inaccurate fault location, high false alarm rate, inaccurate wear threshold prediction, and lack of judgment standards for regular maintenance plan. To solve the above problems, a model of power plant wear prediction and maintenance planning based on particle swarm optimization algorithm is proposed in this paper, and the effectiveness of this method is verified. The algorithm model has low requirement on initial value, no need of cross mutation operation of genetic algorithm, fast convergence speed and easy implementation. Compared with those maintenance plan developed by manual experiences, this algorithm consumes less time, which means the optimal solution will quickly be obtained and costs be reduced. Meanwhile, the maintenance efficiency of the power plant will be greatly improved, providing decision support for the preparation of maintenance plan, pipeline and equipment maintenance procurement, maintenance resource preparation and scheduling plan.