Objective: To determine whether visualization of successful task completion during encoding improves performance on a naturalistic prospective memory (PM) task. Method: The sample was comprised of 60 young adults with HIV infection. Participants were administered a brief clinical neuropsychological assessment, which included a standardized performance-based measure of time- and event-based PM. All participants were also given a naturalistic PM task in which they were asked to manage a mock medication regimen when the examiner showed them the Grooved Pegboard Test during their neuropsychological evaluation. Participants were then randomized into: 1) a visualization condition in which they spent 30 sec imagining successfully completing the naturalistic PM task; or (2) a control condition in which they repeated the task instructions. Results: Multiple linear regression analyses revealed no main effects of visualization or clinical PM for naturalistic PM performance; however, there was a significant interaction between clinical PM ability and visualization condition (p < 0.05). Specifically, for participants with low time-based PM scores, the visualization intervention was associated with significantly higher naturalistic PM accuracy than was the control condition (p < 0.05). In contrast, there was no effect of visualization for participants with high time-based PM scores. There was no interaction between event-based PM and visualization. Conclusion: Findings indicate that a brief visualization exercise can support strategic encoding and performance on naturalistic PM performance in persons with HIV-associated clinical impairments in time-based PM. The extent to which such interventions improve health-related PM outcomes such as medication adherence remains to be determined.