This paper introduces a feedback and feedforward iterative learning control (ILC) scheme for non-affine nonlinear systems featuring iteratively varying trail lengths. The random trail lengths lead to the loss of tracking information in the final iteration. To address this information loss, the deviation in tracking for the ongoing iteration is incorporated with the aid of the feedback control component. It is demonstrated that the convergence condition is contingent solely on the feedforward control gain, with the feedback control part contributing to an acceleration in convergent speed. By establishing the statistical expectation of the initial state as equal to the desired state, it is proven that the mathematical expectation of the error can be effectively controlled to zero. The efficacy of the proposed algorithm is illustrated through numerical simulation.