The current study aimed to investigate prospective memory in traumatic brain injury by using a laboratory-based but naturalistic style assessment task. The objective was to determine if performance differed when the prospective memory targets were self-generated or experimenter-generated (a generation effect). Sixteen people who had sustained a traumatic brain injury and had received rehabilitation were compared to 16 healthy age-matched control participants on a naturalistic virtual measure of shopping and prospective memory. Results demonstrated that the traumatic brain group was less accurate in overall prospective remembering performance than the control group but there was a lack of difference between the self- and experimenter-generated prospective memory targets for both participant groups. Both retrospective memory and executive attention of working memory were associated with prospective memory performance in the naturalistic task. The findings of the study suggest that the generation effect commonly observed in retrospective memory performance may not impact performance in prospective memory.