Attention Deficit/Hyperactivity Disorder (ADHD) is characterised by developmentally inappropriate levels of inattention, impulsivity and hyperactivity. As might be expected of a disorder in which inhibitory deficits form part of the diagnostic criteria, deficits in response inhibition in ADHD have been evidenced in a number of studies. To date, the tasks used in such studies have required participants to inhibit the learned stimulus-response associations that result in unwanted behavior. However, no research has examined the inhibition of stimulus-stimulus associations, formally 'conditioned inhibition'. The present study used video game style conditioned inhibition procedures, developed for children and adolescents with a clinical diagnosis of ADHD and suitable for typically developing matched controls. Two computer-based tasks ('Mission to Mars' and 'Weapon-X') required participants to predict the occurrence of an outcome based on the stimuli presented. We selected 12 male participants with ADHD on medication (methylphenidate), but without comorbid Tourette Syndrome, pervasive developmental disorder, learning disability or psychosis. This group showed overall normal inhibition of stimulus-stimulus associations, measured repeatedly over trials and with two task variants. There was no correlation between inhibitory learning and symptom severity ratings. However, participants with ADHD on higher dosages of methylphenidate, or longer duration of treatment with methylphenidate, showed improved ability to anticipate outcomes following the different stimulus presentations on non-inhibited versus inhibited trials. This effect was most clearly demonstrated on the Weapon-X task. Thus, methylphenidate dose-relatedly improved the expression of associative learning. This action may contribute to its therapeutic effects in improving cognitive function in ADHD.