Boiler is the main equipment of thermal power station. We established a health status evaluation model for the boiler’s superheater through a fuzzy comprehensive hierarchy analysis. Then, we explored the load-damage characteristic curve for ten boilers’ superheaters. Based on this, we aim to minimize the total damage of boilers in the thermal power station through load dispatch. Considering the start-stop state of all the boilers in one plant station, a heuristic Binary Particle Swarm Optimization algorithm for the Unit Commitment load dispatch model is put forward. The algorithm is proved valid to search the feasible particle both in efficiency and effectiveness in the ten-boilers system in 24 h period.