Objectives: The way that individuals perceive themselves and others may be relevant in understanding difficulties with self‐harm. Repertory grids provide a relatively underutilized means of investigating self‐concept in people who self‐harm. This study aimed to examine the association between self‐concept, captured through a repertory grid method, and self‐harm. Method: Ninety‐eight participants with a history of self‐harm took part in the study. A repertory grid was used to derive distances between four elements ('current self', 'ideal self', 'self‐harming self', and 'others who self‐harm'). Associations between these distances and measures of self‐harm behaviour, well‐being and perceived recovery, were estimated. Results: Repertory grid distances were significantly associated with self‐harm. When adjusting for covariates, self‐harm in the past year was positively associated with a greater distance (or less perceived similarity) between current and ideal self, and a smaller distance (greater perceived similarity) between current self and others who self‐harm. Repertory grid distances were also associated with well‐being and perceived recovery. Conclusions: The results are consistent with the idea that self‐perception is associated with self‐harm and related difficulties. The results also highlight the potential utility of repertory grids as a tool for exploring self‐concept. The cross‐sectional nature of the study precludes inferences about directionality or the temporal characteristics of associations. Practitioner points: More recent experiences of self‐harm may be reflected in the way individuals perceive or construe themselves.The way in which individuals who self‐harm perceive or construe themselves may also be related to their current well‐being and perceived recovery.Repertory grids may be a useful tool in clinical practice for investigating self‐concept in those who self‐harm.Repertory grids may provide a more idiographic, clinical outcome assessment in those who self‐harm. [ABSTRACT FROM AUTHOR]