A new P-type iterative learning controller using the concept of the forgetting factor and current error modification is proposed in this paper for robust output tracking of uncertain nonlinear time-varying systems. The robustness, convergence and performance of this P-type learning control scheme are completely studied via a systematic approach. It is shown that uniform boundedness between the system output and the desired output is achieved in each iterate. Under some conditions on the uncertainties, the residual set of the tracking error on the final iterate can be shown to be a class K function of the forgetting factor. When all the uncertainties and the forgetting factor tend to zero, the system output will converge uniformly to the desired one.