The document compares NARMA-L2, model reference, and predictive controllers for a nonlinear quarter car active suspension system. It begins by developing the mathematical model of the suspension system and hydraulic actuator. It then describes the bump and sine pavement road profiles used as inputs. The controller designs are presented, including the NARMA-L2 neural network architecture, model reference design using two neural networks, and predictive design training a neural network to model the plant dynamics. Simulation results show the body travel, acceleration, and suspension deflection outputs for each controller under bump and sine road disturbances. The NARMA-L2 controller performed best in minimizing body travel and acceleration.