This paper introduces Unlearning-Aware Minimization (UAM), a novel min-max optimization framework for effective machine unlearning, which outperforms existing methods on image classification and text-based question answering datasets.