In order to examine how neural network architecture affects the performance of quadcopter control systems, four different PID controllers in Simulink were designed. Three controllers had neural networks and one was a standard PID controller. The response time until the steady-state error reached 0 zero calculated for each controller as a performance metric. After a Chi Square GOF test was conducted based upon the results, it was concluded that there is no significant relationship between the response time and the neural network architecture of the controller.