In this study, the problem of adaptive iterative learning control (AILC) is considered for a class of multiple-input-multiple-output discrete-time nonlinear systems where the initial condition and reference trajectory could be randomly varying in the