<p>Stroke patients experience cognitive impairments affecting daily activities, necessitating effective rehabilitation methods. This study aimed to (1) assess the feasibility and effectiveness of short-term visual memory and selective attention training in an immersive virtual reality environment and (2) examine near and far transfer effects post-training. Twenty-seven stroke survivors were allocated to one of the two groups (Immersive virtual reality – iVR, and Control) and the pre-assessment was completed. Both groups participated in the conventional rehabilitation program. Additionally, the iVR group underwent 10 sessions of short-term visual memory and selective attention training in the iVR environment. After the intervention (or after 2 weeks for the control group) the post-assessment was completed. Cognitive functions were tested by the Addenbrooke’s Cognitive Examination-III and Trail Making Test Part A and B, Medical College of Georgia Complex Figures. The final data analysis included 20 participants (mean age – 62,15 ± 7,8), seven of them did not complete the post-assessment. The results indicated that iVR group (<i>N</i> = 13) showed a significant improvement in short-term visual memory retrieval time, whereas no statistically significant near- or far-transfer effects were observed. Anxiety and depression levels decreased similarly in both groups. In contrast, the control group (<i>N</i> = 7) improved only non-dominant hand psychomotor functions. These findings demonstrate that short-term visual memory and selective attention training tasks, when integrated into an immersive virtual reality environment, may contribute to conventional rehabilitation. However, the present results provide no clear evidence of near- or far-transfer effects following iVR training.</p>