Background: Specific cognitive rehabilitation (SCR) has been suggested for multiple sclerosis (MS). A randomized controlled trial (RCT) evaluating the therapeutic effects of SCR is necessary. Objective: To demonstrate the superiority of a SCR program (REACTIV) over nonspecific intervention (NSI) for neuropsychological (NP) assessment, virtual reality (VR) cognitive testing and daily cognitive functioning. Methods: A single-blind RCT compared SCR and NSI in patients with MS with cognitive complaint. Both programs included 50 individual sessions, 3 times a week for 17 weeks in a real-world setting. The primary end-point was NP assessment. Secondary end-points included semiecological VR tasks (Urban Daily Cog®) and daily cognitive functioning assessment. Maintenance of the effects at 8 months was studied. Results: Of the 35 patients, 18 completed the SCR, and 17 completed the NSI. Several NP and semiecological scores improved significantly more after SCR than after NSI. More NP scores improved significantly after SCR than after NSI. SCR improved daily cognitive functioning. Most improvements were maintained at 8 months. Conclusion: SCR performed in a real-world setting is superior to NSI for improving performance in specific cognitive domains and information processing speed, and for improving cognitive functioning, as evaluated by ecological tools close to daily life and a daily cognitive functioning questionnaire. © 2020 Elsevier B.V.< Réduire