Recently proposed adversarial self-supervised learning methods usually require big batches and long training epochs to extract robust features, which is not friendly in practical application. In this paper, we present a novel adversarial momentum-contrastive learning approach that leverages two memory banks to track the invariant features across different mini-batches. These memory banks can be efficiently incorporated into each iteration and help the network to learn more robust feature representations with smaller batches and far fewer epochs. Furthermore, after fine-tuning on the classification tasks, the proposed approach can meet or exceed the performance of some state-of-the-art supervised baselines on real world datasets. Our code is available at https://github.com/MTandHJ/amoc.