Games are an efficient way to measure the progress of AI. Artificial intelligence has been making rapid advances in recent years, and one way to see how it is progressing is by playing games. Go, Atari2600 and Chess are some AI agents which even defeated human players. These AI agents require or have a training stage involved. In this paper, we propose a training strategy to develop an Artificial Intelligence agent using the NeuroEvolution of Augmenting Topologies (NEAT) algorithm to play the Flappy Bird Game. NEAT basically implements a genetic algorithm that generates ANNs. Our agent learns to safely dodge all the barriers and flap its way through them, optimally playing the famous “Flappy Bird” game.