In this paper, the maximization of aromatics yield is taken as the objective, combined with the kinetic model of photocatalytic reaction, and the reactor inlet temperature is taken as the decision variable. Differential evolution (DE) algorithm is introduced to solve the variable optimally. Aiming at the solution limitations of DE, the adaptive relay-based hybrid differential evolution algorithm is proposed. Combining roulette selection with Gaussian random wandering strategy enhances the ability of global exploration and local exploitation. Through the comparison of algorithm performance and practical application, the improved algorithm in this paper is verified to be effective in enhancing the reaction efficiency of photocatalytic materials. The results show that the unit step deviation of this paper’s algorithm varies in the range of [0.00,0.35]. The optimal value can be basically obtained after about 12 generations of iterations. The difference in the fit of the optimization results does not exceed 0.005. The optimization accuracy is high and the convergence speed is fast. In the photocatalytic chemical simulation experiment, the aromatic yield based on this paper’s algorithm is increased by 3.87%, and the actual aromatic yield reaches 75.89%. Using the algorithm of this paper can improve the optimization effect of photocatalytic reactor, improve the experimental reaction efficiency and increase the aromatic hydrocarbon yield.
