Twenty percent of young people report a lifetime presence of self-harm (SH) behaviour, associated with negative health and functional outcomes. Understanding the underlying cognitive mechanisms is needed to develop targeted early interventions. Reward processing biases may underlie SH, aligning with accounts of the behaviour acquiring “addictive” characteristics. However, the specific nature of such biases remains unclear, particularly its relationship with negative affect (NA) that frequently triggers SH. In Study 1, we compared young people (aged 16–25) with SH to a group with NA but no SH history and a healthy control group on performance of a novel Incentive Delay Task (IDT), with SH-related (SH trials), positive social (social trials) or monetary images (money trials) as stimuli. In Study 2, a different sample of SH and HC participants completed the same IDT following NA induction via an online Trier Social Stress Test. For both studies, we hypothesised faster and more correct responses in the SH group than control groups on SH trials. Contradicting our hypothesis, there were no significant between-group differences in IDT performance on SH, social and money trials in either study. Certain SH characteristics (positive reinforcement, SH mental imagery, urge) were significantly correlated with better performance on SH trials in SH participants. Thus, broadly SH behaviour may not be underpinned by motivational biases towards SH-related cues or naturalistic rewards. Future studies should clarify whether incentivisation of SH-related cues instead explains individual differences in SH behaviour and its relation with treatment and prognosis.