Objective To explore the correlation between physical activity, sleep and aging using a two-sample Mendelian randomization (MR) method. Methods The data through genome-wide association studies was summarized. The single nucleotide polymorphisms (SNPs) related to physical activity and sleep as instrumental variables was selected. The inverse variance weighting method was used for the main analyses, complemented by the weighted median method and MR Egger regression, and then sensitivity analyses were carried out in terms of multiplicity, heterogeneity and leave-one-out method. Finally, multivariate Mendelian methods were applied to eliminate confounders and find mediators. Results A total of two types of physical activity (strong physical activity, physical inactivity) and three sleep conditions (daytime naps, short sleep duration, adequate sleep duration) were found to have a causal relationship with frailty index (P<0.05), while physical inactivity was found to have a causal relationship with telomere length (P<0.05). A total of 167 SNPs were included in the analysis. Strong physical activity [correlation coefficient (β)=−1.26, 95% confidence interval (CI) (−1.60, −0.96), P<0.0001], adequate sleep duration [β=−0.17, 95%CI (−0.26, −0.09), P<0.001] were negatively correlated with the frailty index. Physical inactivity [β=1.47, 95%CI (0.85, 2.08), P<0.001], daytime naps [β=0.25, 95%CI (0.12, 0.39), P=0.0002], and short sleep duration [β=0.20, 95%CI (0.13, 0.27), P<0.0001] were positively associated with frailty index. Physical inactivity [β=−0.38, 95%CI (−0.69, −0.07), P=0.02] was negatively correlated with telomere length. Percentage body fat, body fat mass, waist circumference, body weight and body mass index partially mediated 25.52%, 23.52%, 10.08%, 17.6% and 10.08% of the effect between daytime naps and frailty index, respectively. Conclusion There is a causal relationship between physical activity, sleep, and aging.