Novel robust learning control approach is presented to address nonlinear systems with uncertainties under repeatable control environment. The uncertainties are classified into two categories: time varying parametric uncertainties and norm-bounded nonvanishing uncertainties. The Iterative Learning Control (ILC) scheme is employed to cancel the parametric uncertainties, while the robust control technique is utilized to deal with norm-bounded uncertainties. The concept of Composite Energy Function (CEF) is introduced in the analysis of the learning convergence, consequently the proposed learning control scheme is applicable to quite general systems without requiring the global Lipschitz continuity condition.