Background Late-night overeating (LNOE) is closely associated with many health risk factors, but whether LNOE can increase the risk of death remains unknown. Thus, the prospective cohort study aimed to investigate the relationship between LNOE and mortality using data from the National Health and Nutrition Examination Survey. Methods 11,893 participants aged 50 years and older were included in the study. Dietary information was obtained through 24-h dietary recall interviews. Cox regression, subgroup, sensitivity, and restricted cubic spline analyses were used to assess the association between LNOE and mortality. Results During a median follow-up of 8.3 years, 2,498 deaths occurred. After adjusting for major confounders, compared to the non-late-night eating (NLNE) group, the LNOE group was associated with higher risks of all-cause (HR = 1.47, 95% CI = 1.06–2.04) and cardiovascular disease (CVD) mortality (HR = 2.02, 95% CI = 1.13–3.60). No significant association was found between late-night eating (LNE) and mortality. Subgroup analyses showed that the LNOE group had a greater risk of all-cause and CVD mortality in participants older than 70 years, with alcohol consumption and hypertension and demonstrated an increased risk of all-cause mortality in males and higher CVD mortality in females. Conclusion The habit of LNOE was an independent risk factor for all-cause and CVD mortality in US adults aged 50 years and older, which was also influenced by age, sex, alcohol consumption, and hypertension.