BACKGROUND: Although impaired cognitive control is common during the acute detoxification phase of substance use disorders (SUD) and is considered a major cause of relapse, it remains unclear after prolonged methadone maintenance treatment (MMT). The aim of the present study was to elucidate cognitive control in individuals with heroin use disorder (HUD) after prolonged MMT and its association with previous relapse. METHODS: A total of 63 HUD subjects (41 subjects with previous relapse and 22 non-relapse subjects, mean MMT duration: 12.24 ± 2.92 years) and 31 healthy controls were enrolled in this study. Eye tracking tasks, prospective memory tasks, the Behavior Rating Inventory of Executive Function-Adult Version (BRIEF-A) and the Prospective and Retrospective Memory Questionnaire (PRMQ) were used to assess cognitive control. RESULTS: HUD individuals exhibited worse saccade error rate and executive dysfunction but showed no significant impairment in prospective memory. Additionally, the relapsers performed worse in terms of antisaccade amplitude and velocity at higher difficulty gradients (11° or 16°). Antisaccade performance in terms of amplitude and velocity was negatively correlated with executive function scores. Deficits in inhibition, cognitive flexibility, and self-monitoring were found to mediate the relationship between previous relapse and impaired antisaccade performance. CONCLUSIONS: Even after prolonged MMT, HUD individuals still show partial impairments in cognitive control and antisaccade performance. Previous relapse exacerbates cognitive control deficits through executive dysfunction in inhibition, cognitive flexibility and self-monitoring, which can be screened by higher difficulty of antisaccade amplitude and velocity. More importantly, saccade error rate can reflect impaired inhibitory control in HUD individuals, whereas antisaccade amplitude and velocity appear to have potential diagnostic value for relapse.