Background: Stigma is a barrier to mental healthcare. Understanding context-specific stigma is important in designing mental health interventions. Aim: This study explored the nature and patterns of mental health stigma among community members in Afghanistan. Methods: Using cross-sectional data ( n = 718), descriptive and inferential analyses were conducted. Multivariable linear regression identified determinants of stigma, as reflected by: (1) preference for social distance in relation to depression and psychosis, and (2) intended stigmatising behaviours in relation to mental illness. Results: In descriptive analyses, stigma toward depression and psychosis, especially regarding work and marriage, was found. The label of ‘mental illness’ was especially stigmatising. Most respondents would disclose a mental health problem to family only. Findings of inferential analyses included: stigma was positively associated with higher socioeconomic status, depression and belief in dangerousness; and negatively associated with urban location, female gender, familiarity and belief in a positive prognosis. Conclusions: This study provides new insights on stigma, to inform action in Afghanistan, an under-researched context. Further research and actions must prioritise supportiveness at the family level, especially in remote regions and address cultural issues of the social cost of associating with mental illness. More culturally appropriate, non-stigmatising language surrounding mental illness should be developed.