Abstract: Background　High-altitude hypoxia and environmental stress significantly impair core cognitive functions such as attention, memory, and executive control, thereby compromising overall brain efficiency and operational performance. Traditional paper-and-pencil assessments have limited ecological validity and are insufficient for monitoring cognitive function in complex environments. Objective　To evaluate the sensitivity and applied value of a virtual reality (VR)-based cognitive assessment system in detecting cognitive changes under high-altitude exposure. Methods　Healthy male young adults scheduled to enter a highaltitude region were recruited. Cognitive assessments were conducted at plain land (approximately 300 m above sea level) and one week after arrival at the high-altitude region (approximately 3 650 m above sea level). The Montreal Cognitive Assessment (MoCA), the Stroop test, and the Digit Span test were administered as traditional assessment tools, while VR-based tasks including target discrimination, map navigation, sound localization, and simulated operation were concurrently performed. The standardized response mean (SRM) and Cohen's d_av were calculated across cognitive dimensions to compare the magnitudes of change detected by the traditional assessments and VR tasks across different cognitive domains. Spearman correlation analysis was used to examine the associations between VR task performance and traditional measures, and the intraclass correlation coefficient (Intraclass Correlation Coefficient ICC) (2,1)was employed to evaluate the stability of inter-individual differences in VR indicators between plain and high-altitude assessments before and after high-altitude exposure. Results　A total of 200 healthy young males with a mean age of (21.4±7.5) years were recruited. After 1 week at high altitude, the participants exhibited a significant decline in MoCA scores (SRM=−0.570, P＜ 0.001). Significant changes were observed in multiple cognitive dimensions common to both traditional assessments and VR indicators. For attentional accuracy, the effect size of VR target discrimination accuracy (SRM=−1.736) notably exceeded that of Stroop accuracy (SRM=−0.760); for working memory, VR map navigation time (SRM=0.963) showed a markedly larger effect than backward digit span (SRM=−0.422); for spatial localization and interference inhibition, VR sound localization angular deviation (SRM=1.035) and response time (SRM=0.769) both demonstrated large effect sizes. Spearman correlations revealed that VR map navigation time was negatively correlated with backward digit span (r=-0.452, P＜0.001), while VR target discrimination accuracy was positively correlated with Stroop accuracy (r=0.232, P＜0.01). ICC analyses indicated that processingspeed- oriented VR indicators, such as target discrimination time (ICC=0.787) and simulated operation reaction time (ICC=0.750), exhibited good stability of inter-individual differences between plain and high-altitude assessments. Conclusion　High-altitude exposure leads to cognitive decline in young adults. The VR-based cognitive assessment system can sensitively capture such changes, with core indicators demonstrating satisfactory validity, stability, and applied value.