One main problem faced by people with multiple sclerosis (PwMS) is upper limb dysfunction, which can occur in the first decade of the disease and with the highest prevalence of disability in the progressive type of the disease. Then, PwMS may benefit from personalised and intensive treatment as provided by robotic devices. These innovative devices have increasingly been brought into the neurorehabilitation field, due to their ability to provide repetitive and task-oriented training. In this quasi-randomized study, we aim to evaluate the effects of robotic-assisted hand training, using the Hand TutorTM device, on hand functionality, active RoM, and manual dexterity, compared to conventional rehabilitation in PwMS. We enrolled 30 MS patients, who received 20 training sessions, each lasting 45 min with robotic-assisted hand training with Hand Tutor (n 15, experimental group) or conventional rehabilitation therapy (n 15, control group). All patients were evaluated at pre- and post-intervention with clinical scales for upper limb functionality (DASH, BBT, NHPT, and MI). In addition, only patients in the experimental group received an objective kinematic analysis of the hand and wrist movements, delivered by the Hand Tutor glove, both pre- and post-intervention. We found that PwMS in both groups statistically improved their upper limb functions, however the experimental group achieved better results in terms of manual dexterity. This promising rehabilitation training with Hand Tutor glove led to positive effects on upper limbs motor outcomes and kinematic parameters in patients with MS.