A performance optimization model of direct-contact heat exchangers was established, in which the volumetric heat transfer coefficient was used as objective function, the refrigerant initial flow rate U₀, the nozzle diameter d_i, and the heat transfer oil height Z were used as decision variables. The linear model of bubble swarm and heat transfer performance was used as constraint condition in the optimization model, and the influence of the constraint condition on optimization process and results were mainly analyzed. The genetic algorithm was used to analyze the original model and supplementary model. And the results show that the optimized volumetric heat transfer coefficient of the original model is promoted by 6.7 times compared with that of the original design. The number of iterations of the optimal value in the supplement model is reduced by around 55%, and the optimal value increases by 0.3% compared with those of the original model. Therefore, this constraint condition improves not only the iteration rate, but also the probability of seeking the global optimum, which makes the optimal solution more approximate to the global optimal value.