BACKGROUND: Whether sedentary behavior and physical activity causally influence hypertension remains uncertain. We used Mendelian randomization (MR) to examine whether genetically predicted domain-specific sedentary behaviors and leisure-time physical activity are associated with hypertension risk and to explore potential cardiometabolic pathways. METHODS: We performed a two-sample MR using genome-wide significant, LD-independent SNPs as instruments for domain-specific sedentary traits and physical activity. Hypertension was defined by the FinnGen R9 endpoint I9_HYPTENSESS. Pleiotropy-robust estimators, outlier procedures and two-step MR mediation analyses were applied. RESULTS: Genetically proxied leisure screen time (LST) was associated with a higher risk of hypertension (IVW OR = 1.181, 95% CI 1.076-1.296, p = 4.70e-4). Two-step MR suggested that body mass index (BMI) and type 2 diabetes (T2D) may partly account for this association, with estimated mediated proportions of 54.7% and 14.7%, respectively, although these estimates should be interpreted cautiously. No robust associations were observed for moderate-to-vigorous physical activity (MVPA) or sedentary behavior at work (SDW). The MVPA estimate was directionally protective, but these null findings remained inconclusive given lower instrument informativeness. CONCLUSION: Genetically proxied LST was associated with higher hypertension risk and may represent a plausible, potentially actionable prevention target. BMI and T2D may partly account for this association. No robust associations were observed for MVPA or SDW, but these null findings should not be interpreted as definitive evidence of no effect. Further validation in prospective, multi-ancestry, and mechanistic studies is needed.