In this paper, the two iterative learning control methods which improve path tracking performance of robot are discussed. In these methods, the parameters of robot are estimated from simple robot model, and then iterative control input which is composed of feedforward part and feedback part is applied. The feedforward part is parameter estimated nonlinear robot dynamics term and iterative learning input term. The feedback part is position error and velocity error term. The first method is proposed to get good performance in less trials and the second method is proposed to get better performance when the learning results on one path are used on new path than they are not used. From the simulation, the first method shows good path tracking performance in less trials and second method shows better path tracking performance when learning results are used on new path.