Abstract Introduction Manganese (Mn) is an essential trace mineral required for normal growth and physiological processes. However, inadequate or excess Mn can have a detrimental impact on human health. Despite the essentiality, little is known about the relationship between Mn and sleep. The objective of this study is to examine the relationships between blood Mn levels and sleep patterns (duration, midpoint of sleep, obstructive sleep apnea [OSA] symptoms, and trouble sleeping) in U.S. adults. Methods Data from the 2017 to 2020 National Health and Nutrition Examination Survey (NHANES) was used for the analysis. Adults aged 18 or older with both blood Mn levels and sleep parameters were included (n=8356). Self-reported fall asleep time and wake up time in the past week were used to calculate sleep duration and midpoint. Trouble sleeping and symptoms of OSA were also collected. Blood Mn levels were divided into quintiles for analysis. Multivariable logistic regressions that accounted for complex study design were used. Results The mean (SE) blood Mn concentration was 9.7 (0.1) µg/L, and was higher among participants who were younger, female, identified as non-Hispanic Asian, had lower income, were non-smokers, and had lower iron levels (P< 0.05, for all). About 25% of adults had short sleep duration (< 7 hours) on weekdays. OSA symptoms were noted in 49% of adults. In men, the odds of short sleep duration on weekdays were higher among those with the lowest Mn level (Q1: < 7.0 µg/L) than the reference group (Q3: 8.5-10.0 µg/L) after adjusting for age, race, income, smoking, and iron (OR: 1.5, 95% CI: 1.1-2.1, P=0.03). Among women, Q2 (7.0-8.4 µg/L) was the related to the lowest odds of OSA symptoms compared to all other groups after adjusting for covariates (P=0.049). Conclusion Lower or higher Mn levels in blood were related to poor sleep health in U.S. adults, which suggests that inadequate or excessive Mn intake might be harmful to sleep. Future studies on the mechanisms of these relationships are needed. Support (if any) This work was supported by National Institutes of Health R01DK123022, R21NS112974 to YAS, and K01HL151673 to ECJ.