Prospective Memory (PM) is the ability to remember to do something in the future. It is often impaired after stroke and can impact on an individual’s level of independence and daily functioning. PM tasks have been criticized for their lack of ecological validity wherein test results may not be related to actual performance in daily life. With ecological validity in mind, the Virtual Reality Prospective Memory Shopping Task (VRPMST) was designed to assess two types of PM, time- and event-based. This study aimed to examine the ecological and convergent validity of the VRPMST in comparison to an experimental (Lexical Decision PM Task) and clinical measure of PM (Cambridge PM Test). Twelve individuals with stroke and 12 controls were administered three PM measures, three neuropsychological measures, and two user-friendliness questionnaires, one for the experimental PM measure and one for the VRPMST. Individuals with stroke showed impairments in PM compared to controls on all three PM measures, particularly time-based PM. Individuals with stroke were found to monitor time significantly less than controls on both the experimental PM measure and the VRPMST. The VRPMST was found to be sensitive in measuring PM, have better ecological validity when compared to the experimental PM measure, and good convergent validity. The findings of this study have helped to clarify that PM impairment does exist after stroke, possibly due to a problem in strategic monitoring. In addition, we have demonstrated how VR technology can be used to design a measure of cognitive function commonly impaired after stroke.