Our aim was to determine if deficits in intentional forgetting that are associated with depression and dysphoria (subclinical depression) could be explained, at least in part, by variations in working memory function. Sixty dysphoric and 61 non-dysphoric participants completed a modified version of the think/no-think (TNT) task and a measure of complex working memory (the operation span task). The TNT task involved participants learning a series of emotional cue–target word pairs, before being presented with a subset of the cues and asked to either recall the associated target (think) or to prevent it from coming to mind (no think) by thinking about a substitute target word. Participants were subsequently asked to recall the targets to all cues (regardless of previous recall instructions). As expected, after controlling for anxiety, we found that dysphoric individuals exhibited impaired forgetting relative to the non-dysphoric participants. Also as expected, we found that superior working memory function was associated with more successful forgetting. Critically, in the dysphoric group, we found that working memory mediated the effect of depression on intentional forgetting. That is, depression influenced forgetting indirectly via its effect on working memory. However, under conditions of repeated suppression, there was also a direct effect of depression on forgetting. These findings represent an important development in the understanding of impaired forgetting in depression and also suggest that working memory training might be a viable intervention for improving the ability of depressed individuals to prevent unwanted memories from coming to mind.