Abstract: In order to comprehensively evaluate the generating capacity of generator units, the K\|means PSO clustering algorithm is presented. This algorithm combines the features of K\|means clustering algorithm and particle swarm optimization, which can quickly classify data sets and has a better convergence effect. This method is used to comprehensively evaluate 10 generator units, evaluation value for each generator unit is obtained through clustering analysis, calculation of the weights and importance degree, and the values are sorted by their sizes. The results verify the effectiveness and feasibility of the method proposed in this paper, which will provide the necessary analysis basis for energy-saving power generation, and has certain practical significance.