The ability to remember and perform delayed intentions was investigated in a sample of 40 participants with Multiple Sclerosis (MS) and compared to a group of 36 healthy, neurologically intact, control participants. Using a model of task division, performance on both the prospective memory (PM) component (a failure to remember to do something at a specific time) and the retrospective memory (RM) component (a failure to remember the content of the intention) of two "naturalistic" delayed intention experimental tasks were examined. Significantly poorer performance of the MS group in completing one of the delayed intention tasks successfully appeared to be primarily attributable to retrospective memory deficits rather than prospective memory deficits. This proposition was further supported by group differences on RAVLT measures of retrospective memory. By utilising a paradigm that enables the nature of failure in performing delayed intention tasks to be identified, specific strategies for the clinical management of MS cognitive deficits can be developed.