This article studies the feasibility and effectiveness of the fusion model of cultural algorithm (CA) and projection pursuit (PP) in similar watershed optimization . Based on the optimization of the control catchment of the 12 river stations, this article presents a CA-PP similarity basin optimization model and constructs a differential evolution (DE) algorithm-PP, a harmony search (HS) algorithm-PP and a particle swarm optimization (PSO) algorithm-PP as a contrast model, the optimal results are compared with the optimal results of stochastic analysis, set pair analysis, fuzzy analysis and gray analysis. The results show that the optimal value, the worst value, the mean value and the standard deviation obtained by CA-PP are better than those of the DE, HS and PSO algorithms, and they have better global optimal value and convergence stability. The result of CA-PP model is similar to that of DE-PP, HS-PP and PSO-PP, and the results of random analysis, set pair analysis, fuzzy analysis and gray analysis are the same, but there are differences in the optimization sequence. The CA-PP model is feasible and effective for similar watershed optimization, and it can provide new approaches and methods for similar optimization.