Decanter centrifuges are widely used for solid–liquid separation. Although parameter analysis for decanter centrifuges was performed by numerical simulation in previous studies, some structural parameters are rarely mentioned and investigated. At the same time, the results obtained by the single-parameter analysis in previous studies are difficult to truly realize the comprehensive performance optimization of decanter centrifuges. In this paper, the influences of the window structure and bowl–conveyor gap on the separation performance are systematically analyzed with the employment of a numerical computation method. The results show that the increase in the window angle and window height will accelerate the flow of the upper layer, while the increase in the bowl–conveyor gap may make particles flow through it directly and further form a solid retention zone. Both of the structural changes will lead to deterioration of the separation performance. On the basis of numerical simulation analysis, a genetic algorithm-based method for multiparameter optimization is proposed in this paper. Parameter optimization shows that bowl speed and feed flow rate have the most significant effects on the separation performance and power consumption. Compared with the minimal specific power in the first generation, the optimized specific power is reduced by 15.7%, and the cake solid content merely decreases by 0.044%.