Background: Sleep duration (SD) is a critical determinant of health, with both short and long SD associated with increased risks of morbidity and mortality. However, distinct ages may exhibit unequal sleep needs. Few studies have addressed the modulation of age on SD associated all-cause mortality (AM). This study aimed to investigate age- and gender-specific associations between SD and AM. Methods: Four national cohort studies were conducted: National Health and Nutrition Examination Survey (NHANES), Survey of Health, Ageing and Retirement in Europe (SHARE), Chinese Longitudinal Healthy Longevity Survey (CLHLS), and China Health and Retirement Longitudinal Study (CHARLS). Data pre-processing such as integration, weighting and interpolation was conducted following cohort-specific user manuals. Restricted cubic spline (RCS) was used to test the nonlinear relationship. Kaplan-Meier and Cox regression survival analysis were used to evaluate the association between SD and AM. The moderation of mortality by SD was subsequently investigated in adult, middle-age, elderly and longevity. Results: RCS analysis showed a nonlinear correlation between SD and AM in the four cohorts(P < 0.001). Survival analysis reveals 6-8 h to be the optimal SD. Subgroup analyses showed a higher risk in adults with <= 5 h than >9 h (HR: 2.95 [1.90-4.59] vs 1.83 [1.08-3.12]), and a similar risk in middle age (HR: 2.42 [1.76-3.32] vs 2.44 [1.64-3.61]). In contrast, four datasets suggest a higher risk of the elderly being affected by long SD. Adults (HR = 1.82 [1.14-2.92], P = 0.012) and middle age (NHANES: HR = 1.50 [1.04-2.16], P = 0.029; SHARE: HR = 3.04 [1.10-8.41], P = 0.032) significantly increased the risk of short sleep compared to elderly. Conclusions: SD associated mortality risk varies with age. Young adults should avoid poor sleep, while elderly should avoid sleeping too much.