The crosslinked polymer flooding, which is developed on the basis of polymer flooding, is a new type of flooding technology. As an EOR method, cross-linked polymer flooding has become a research hotspot. In the process of cross-linked polymer flooding, if the concentrations of the polymer and the crosslinking agent are small, the viscosity of the solution is low, and it will not achieve the oil displacement effect. Meanwhile, if the concentrations of the polymer and crosslinking agent are large, the viscosity of the solution is high, it needs high pressure to drive it flowing in the formation. Further, with the increasing injection of chemical agents, the contradiction between reduced production and increased cost has presented. The performance of crosslinked polymer flooding depends on the interaction of these two factors. Therefore, the concentrations of polymer and crosslinking agent should be optimized. In this paper, an optimal design method is proposed by using genetic algorithm with global optimization characteristics algorithm, combining with the chemical flooding numerical simulation software UTCHEM, the concentrations of the chemical agents are optimized. Firstly, the cumulative oil production is calculated by numerical simulation software UTCHEM, then the concentrations of the chemical agents are randomly generated by the genetic algorithm in the encoding process, and the fitness function takes the profit of cross-linked polymer flooding. Given a set of initial values, through crossover and mutation of population, optimized injection concentrations of the polymer and cross-linking agent are obtained by the multi-generational calculation.