Evolutionary techniques are useful for processing large volumes of information for an effective decision-making; their versatility allows them to be applicable in various environments. In software development it is important the continuous improvement of the processes and the reuse of experiences to guide organizations evolution efforts. This work develops a genetic algorithm, which redefines the selection, crossing and mutation operators to solve scenario optimization problems in Software Process Improvement. The objective function of the genetic algorithm integrates the implementation of an evolutionary artificial neural network to predict the success of the scenarios, as well as the use of association rules to identify dependencies between Good Practices and Critical Success Factors in Software Process Improvement. The validation of the solution, through the Iadov technique and a quasi-experiment with multiple chronological series, showed that scenario optimization contributes to decision-making in Software Process Improvement.