A prospective cohort study estimated the incidence of hepatitis C virus (HCV) in drug injectors in South Wales (UK). In total, 286/481 eligible seronegative individuals were followed up after approximately 12 months. Dried blood spot samples were collected and tested for anti-HCV antibody and behavioural data were collected at baseline and follow-up. HCV incidence was 5·9/100 person-years [95% confidence interval (CI) 3·4–9·5]. HCV incidence was predicted by community size [incident rate ratio (IRR) 6·6, 95% CI 2·11–20·51, P=0·001], homelessness (IRR 2·9, 95% CI 1·02–8·28, P=0·047) and sharing injecting equipment (IRR 12·7, 95% CI 1·62–99·6, P=0·015). HCV incidence was reduced in individuals in opiate substitution treatment (IRR 0·34, 95% CI 0·12–0·99, P=0·047). In order to reduce follow-up bias we used multiple imputation of missing data using switching regression; after imputation estimated HCV incidence was 8·5/100 person-years (95% CI 5·4–12·7). HCV incidence varies with community size, equipment sharing and homelessness are associated with increased HCV incidence and opiate substitution treatment may be protective against HCV.