Starting from the economic dispatch problem of power system unit combination, this paper analyzes the high-dimensional, discrete and nonlinear mathematical model of unit combination model, using the robust optimization theory to study the unit combination mathematical model and solution strategy, and based on the existing problems of the light robust model, an improvement study was carried out, and a new improved light robust optimization model was proposed. The algorithm for solving the economic distribution model is designed, and the binary particle swarm optimization algorithm is improved by using the approximate random search and ergodic characteristics of chaos-simulation. Chaos-simulated annealing improved binary particle swarm optimization can optimize the search speed of the optimal solution and avoid simulated annealing failure, and improve the global search ability of the algorithm. The improved particle swarm algorithm is used to solve the economic distribution model of the unit combination. The speed difference between the individual and the group in the algorithm updates the particle speed, which improves the global search ability and later local search ability of the particle algorithm. Combined with a specific power system unit combination example, the established unit combination load economic distribution model and solution algorithm are verified. The results show the effectiveness and superiority of the designed load economic distribution model and the improved particle swarm algorithm.