Although emerging studies from high-income countries investigated the relationship between sedentary behavior (SB) and cardiovascular risk, little evidence came from developing countries. Moreover, the benefits of reallocating time from SB to physical activity (PA) on incident cardiovascular disease (CVD) are unknown. Using three cohorts from the Prediction for Atherosclerotic Cardiovascular Disease Risk in China project, we included 93 110 adults who were free from CVD at baseline. Cox proportional hazards models were used to calculate the hazard ratios (HRs) and 95% confidence intervals (CIs) for CVD, including stroke, coronary heart disease, and CVD death. Isotemporal substitution models were applied to estimate the per-hour effects of replacing SB with PA. After 5.8 years follow-up, 3799 CVD cases were identified. A gradient positive association between sedentary time and incident CVD was observed. Relative to those with < 5 h/d sedentary time, the multivariable-adjusted HRs (95% CIs) of CVD incidence were 1.07(0.96–1.20), 1.27(1.13–1.43) and 1.51(1.34–1.70) for those having 5–<8, 8–<10, and ≥ 10 h/d sedentary time, respectively. When participants were cross-classified by SB and moderate to vigorous physical activity (MVPA) level, the CVD risk was highest in those with ≥ 10 h/d SB and < 150 min/week MVPA. Among those who reported ≥ 5 h/d sedentary time, per-hour substitution of SB with light, moderate, and vigorous PA reduced incident CVD risk by 5%, 6%, and 8%, respectively. The study first found that sedentary time was associated with increased incident CVD risk among Chinese adults and that substitution of SB with PA of any intensity could convey cardiovascular benefits among those with ≥ 5 h/d SB.