BACKGROUND: Calcium is involved in many biological processes, but the impact of serum calcium levels on long-term mortality in general populations has been rarely investigated. METHODS: This prospective cohort study analyzed data from the National Health and Nutrition Examination Survey (1999-2018). All-cause mortality, cardiovascular disease (CVD) mortality, and cancer mortality were obtained through linkage to the National Death Index. Survey-weighted multivariate Cox regression was performed to compute hazard ratios (HRs) and 95% confidential intervals (CIs) for the associations of calcium levels with risks of mortality. Restricted cubic spline analyses were performed to examine the non-linear association of calcium levels with all-cause and disease-specific mortality. RESULTS: A total of 51,042 individuals were included in the current study. During an average of 9.7 years of follow-up, 7,592 all-cause deaths were identified, including 2,391 CVD deaths and 1,641 cancer deaths. Compared with participants in the first quartile (Q1) of serum calcium level [≤2.299 mmol/L], the risk of all-cause mortality was lower for participants in the second quartile (Q2) [2.300-2.349 mmol/L], the third quartile (Q3) [2.350-2.424 mmol/L] and the fourth quartile (Q4) [≥2.425 mmol/L] with multivariable-adjusted HRs of 0.81 (95% CI, 0.74-0.88), 0.78 (95% CI, 0.71-0.86), and 0.80 (95% CI, 0.73, 0.88). Similar associations were observed for CVD mortality, with HRs of 0.82 (95% CI, 0.71-0.95), 0.87 (95% CI, 0.74-1.02), and 0.83 (95% CI, 0.72, 0.97) in Q2-Q4 quartile. Furthermore, the L-shaped non-linear associations were detected for serum calcium with the risk of all-cause mortality. Below the median of 2.350 mmol/L, per 0.1 mmol/L higher serum calcium was associated with a 24% lower risk of all-cause mortality (HR: 0.76, 95% CI, 0.70-0.83), however, no significant changes were observed when serum calcium was above the median. Similar L-shaped associations were detected for serum calcium with the risk of CVD mortality with a 25% reduction in the risk of CVD death per 0.1 mmol/L higher serum calcium below the median (HR: 0.75, 95% CI, 0.65-0.86). CONCLUSION: L-shaped associations of serum calcium with all-cause and CVD mortality were observed in US adults, and hypocalcemia was associated with a higher risk of all-cause mortality and CVD mortality.