Children's performance on tests of visuo-spatial working memory improves with age, although relatively little is known about why this happens. One explanation concerns the development of the ability to recode visually presented information into phonological form. This process appears to be used from around 8 years of age and is a major contributor to tasks in which stimuli can be verbally labelled. However, evidence suggests that phonological recoding cannot account for all of the age-related change in performance on visuo-spatial working memory tasks. In this review, four other mechanisms (knowledge, processing strategies, processing speed, and attentional capacity) are considered in terms of their contribution to children's visuo-spatial working memory development.