The use of neuro-evolution techniques for intelligent agents functioning in dynamic complex environments poses a unique opportunity for artificial intelligence researchers. Training AI agents against human players is difficult due to the substantial amount of time and exposure needed to adequately develop the agents. Therefore, it is imperative that we develop and evaluate methods of training agents against each other. This raises a separate challenge to regular machine learning problems, using co-evolution with competitive populations where the environment of learning and fitness of an agent is in an evolving landscape. This project explores using Neuro-evolution of augmenting topologies (NEAT) algorithm in two separate co-evolving populations where one populations actions directly influences the outcomes of the other and assesses its effectiveness when used with a static fitness function.