Plant breeders widely use recurrent selection schemes to increase the frequency of favorable alleles for quantitative traits in a population. Although simultaneous selection is complex because it involves several traits combined with selection cycles, the use of selection indexes (SI) is applied to increase the chance of success of the breeding program. As many indices are available in the literature, simulations can help breeders to determine which selection index can be better adjusted considering the selection goals, the intensity, and the genetic correlation among traits over breeding cycles. In this context, we estimated the realized genetic gains in a tropical maize haploid inducer population after two cycles of recurrent selection, using external testers and optimizing the simultaneous selection for this breeding population in the long-term via stochastic simulations. Furthermore, we proposed a new approach to optimize the initial weights by applying the Smith-Hazel method to maximize the genetic gains for all traits in a balanced way. Overall, the estimated gains in real induction rate were about 63% per cycle, improving the population performance from 0.8 to 2.8%. Moreover, our results confirm that the traditional Smith-Hazel approach outperformed other methods for long-term response to selection. Finally, recurrent selection with external testers may be a suitable method to improve the haploid induction rate in tropical maize populations.