Objectives <p>Observational studies have consistently shown physical activity associated with lower mortality. Randomized controlled trials to confirmAuthors’ contributions the value of physical activity for lifespan in the general population are challenging to conduct. To address this gap, we conducted a Mendelian Randomization (MR) study, using the largest available suitable studies and control outcomes.</p> Method <p>We conducted a two-sample MR study using summary statistics in European populations. We used single nucleotide polymorphisms (SNPs) strongly (<i>p</i> &lt; 5 × 10^− 8), and independently (r² &lt; 0.001) predicting leisure-time moderate to vigorous intensity physical activity (<i>N</i> = 608,595) and SNPs similarly predicting inactivity (<i>N</i> = 526,725, leisure-time screen time) to obtain inverse variance weighted estimates. Lifespan was proxied by parental attained age (current age or age at death). We adjusted for education using multivariable MR. Waist circumference and whole-body fat mass were control outcomes. Sensitivity analyses included the weighted median, MR-Egger, MR-PRESSO and MRlap.</p> Results <p>Leisure-time moderate to vigorous intensity physical activity was associated with longer lifespan for men (1.41 years per doubling the odds of being physically active in leisure time, 95% confidence interval (CI) 0.21 to 2.62) and women (1.68 years, 95% CI 0.12 to 3.25). Conversely, leisure-time screen time was associated with shorter lifespan, which was less evident after adjusting for education. As expected, leisure-time moderate to vigorous intensity physical activity reduced and screen time increased waist circumference and whole-body fat mass, which remained after adjusting for education.</p> Conclusions <p>Leisure-time moderate to vigorous intensity physical activity may extend lifespan, while the role of leisure-time screen time is less clear. Questions remain about the optimal type, duration, intensity, and frequency of physical activity.</p>