Background. Stroke is a leading cause of permanent disability. Being a new technology for neurorehabilitation, virtual reality (VR) allows for intensive trainings with a larger number of repetitions focused on mastering specific skills in patients with upper limb dysfunction. To date, there are insufficient studies evaluating the use of VR in the acute period of a stroke. Objective. To investigate the effectiveness and safety of adding VR trainings to standard therapy for improving upper limb function and for enhancing activities in daily life in patients with acute stroke. Subjects and method. The investigation enrolled 78 patients with acute stroke (the median time from stroke onset, 3.7 days; median age, 63 (43; 79.3) years), who were randomized into 2 groups: a study group (standard therapy + VR) and a control one (standard therapy only). The patients of the study group underwent a VR training cycle lasting 15 minutes twice daily for 10 days. Before and after the training cycle, the British Medical Research Council scale for muscle strength, the upper extremity motor function sections of the Fugl—Meyer assessment (FMA) scale, hand dynamometry, and the nine hole peg test (NHPT) were used to evaluate upper extremity function in the participants; Sections 6 and 7 of the motor function assessment scale (MAS), the modified Rankin scale (MRS), and the total motor function scores of the functional independence measure (FIM) utilizing a 7-point ordinal scale were employed to assess everyday activity limitations. Results. On completion of rehabilitation measures, the participants in the study and control groups showed a statistically significant improvement in upper extremity function and activities of daily living. According to the ratings of the FMA subscales «hand» (p=0.034), «hand (speed)» (p<0,001), and the total score of this scale (p=0.035) and according to those of the FIM scale (p=0.045), the patients of the study group demonstrated a statistically significant improvement compared with those in the control group. Assessments using the MRC scale, FMA subscales «upper extremity» and «wrist», NHPT, MAS, the Barthel index (BI), and MRS revealed no differences between the groups. No serious adverse events were recorded in the participants of the study group during the study period. Discussion. The vast majority of studies evaluating the effect of VR in upper extremity dysfunction are conducted in patients with stroke lasting longer than 3 months. Despite the validity of using VR in the acute period of cerebral lesion, the investigations of this design are not numerous. At the same time, these studies often have limitations, such as a small number of participants (usually less than 10); a small number of training VR sessions; lack of «blinding»; different types of software and hardware systems that implement VR technology; heterogeneity of participants according to the degree of upper extremity dysfunction; insufficient detailing of the content and scope of conventional therapy; comparison of VR with basic methods of motor rehabilitation, and sometimes the absence of a control group. Conclusion. Thus, the use of VR in addition to standard rehabilitation measures in the acute period of stroke favors improvements in upper extremity function and a reduction in daily limitations and is safe.