A review of the literature on human and non-human animal self-injury reveals that there has been little cross-fertilization of ideas between these two domains. Each body of research offers particular strengths, which, if combined, may more keenly inform clinicians and researchers alike about the nature of human self-injury. While research on self-injury among humans has necessarily relied upon correlational designs, researchers who study self-injury among non-human primates have more tightly controlled experimental methods at their disposal. Experimental research allows researchers and clinicians to address issues of causality in ways that are not possible with correlational designs. Despite the difficulties of drawing inferences from different species, a review of the experimental data on non-human primate self-injury may give researchers fresh insights into the elusive nature of human self-injury. This article highlights the research and theoretical material pertaining self-injury and identifies ways in which animal models of self-injury might inform research and clinical understanding of human self-injury.