Background: Multiple sclerosis (MS) can have dramatic effects on performance of everyday life activity. However, the ability to assess everyday functional activity remains elusive. The purpose of this study was to establish validity and reliability to a performance-based assessment of everyday life activities called actual reality (AR). AR involves utilization of the Internet to perform three actual everyday life activities: purchasing (a) an airline ticket, (b) cookies, and (c) pizza. Method: A repeated measure design was used to examine 30 adults with MS who were recruited from a nonprofit rehabilitation research institution and 30 healthy controls (HC) living in the community. Participants were administered the 3 AR tasks twice, 3 weeks apart. Additionally, neuropsychological tests and self-report functional questionnaires were administered. Results: This study supported moderate to large Interrater Reliability of the AR assessment. Additionally, the 3 AR tasks did not differ in the number of errors made, and number and quality of cues required to complete the AR tasks. Participants with MS committed more errors, and required significantly more cues to perform the 3 AR tasks successfully compared with HC, supporting discriminant validity of the AR. Concurrent validity was supported by moderate to large associations between AR performance and neuropsychological test scores. Practice effect was observed for cognitive processes and the time it took participants to perform the task. Conclusions: Data support the use of actual, real-life, performance-based approach to measuring functional cognition outcomes. However, the observed practice effects for AR-cognitive capacity and AR-latency should be noted.