Recent years have witnessed an increasing number of empirical studies in second language acquisition (SLA) pointing to the important role of working memory (WM) in various aspects of L2 learning. Major findings of this body of research are outlined here, which is followed by a summary of potential limitations and caveats in research design and methodology. To address these issues, we draw on nomothetic WM theories and WM-SLA research insights to propose a theoretical model to conceptualize and measure the WM construct in specific SLA domains and activities. Basic tenets and principles of this integrated WM-SLA model are highlighted as well as their underlying rationale. Overall, it is argued here that such an integrated theory makes vital contribution to the advancement of WM-SLA research and practice.