This research tests the role of visual perspective taking (VPT) in mediating the relation between spatial ability and theory of mind (ToM). Study 1 demonstrated such mediation correlationally in seventy 3.5‐ to 4‐year olds. In Study 2, twenty‐three 3.5‐ to 4‐year‐olds were trained on using play blocks to copy preassembled models as a way to promote spatial ability. Resultant increases in VPT and ToM were compared to those from a control group learning to draw instead (n = 23). Both studies showed that the effect of spatial ability on ToM depended on VPT, suggesting a role of embodiment in ToM development in early childhood. These findings provide an alternative way to think about ToM development and the psychological mechanism that may be involved. Copyright © 2021 Society for Research in Child Development.