This dissertation, submitted in partial fulfillment of the PhD degree in Computer Science, focuses on improving the performance of the NEAT neuroevolutionary algorithm by accelerating the fitness evaluation stage through distributed parallel computing. Since evaluating each neural network in a population is computationally expensive and traditionally performed sequentially, it becomes the main performance bottleneck. The proposed approach enables parallel execution of fitness evaluations across distributed computing resources, significantly reducing computation time, improving scalability, and extending the applicability of neuroevolutionary algorithms to complex real-world tasks, including autonomous robotics and unmanned aerial vehicles.