This work explores the use of the NEAT (NeuroEvolution of Augmenting Topologies) algorithm to train artificial neural networks for autonomous driving in a simulated environment built with Pygame. The virtual vehicle perceives its surroundings through distance sensors and learns to navigate without prior human data. The NEAT algorithm evolves both the structure and weights of the networks through generations. The simulation tests include varying track difficulties and obstacle scenarios. Results show that the AI can learn effective navigation strategies over time. NEAT proves to be a viable approach for autonomous control where supervised data is unavailable.