The focus of this article is the well documented association between low working memory capacity and difficulty with mathematical word-problem solving. We begin by describing a model that specifies how various cognitive resources, including working memory, contribute to individual differences in word-problem solving and by then summarizing findings on the relation between working memory and word-problem solving. This sets the context for the article's main purpose and major section: to describe the findings of research studies that take one of two approaches for addressing the needs of students with low working memory within word-problem solving intervention. One approach focuses on "compensating" for working memory limitations; the other on "building" working memory capacity. We then suggest the need for research on integrating the two approaches by embedding working memory training within explicit word-problem solving intervention. [This is the in press version of an article published in "ZDM Mathematics Education."]