Introduction: Chronic disruption of circadian rhythms is linked to weight gain and metabolic dysregulation. Time-restricted eating (TRE), a form of intermittent fasting, has shown effectiveness in improving short-term weight loss and energy homeostasis. However, associations between habitual adherence to TRE and long-term weight change remain understudied. Hypothesis: We tested whether adherence to TRE assessed by the gold-standard seven-day dietary records (7DDRs) was related to 5-year change in body weight among women. Methods: The present analysis included 650 women (mean [SD]: age 63 [9] y; BMI 26.4 [5.3] kg/m 2 ) without cardiovascular disease at baseline who completed 7DDRs in a sub-study of the Nurses’ Health Studies (NHS/NHSII), the Women’s Lifestyle Validation Study (WLVS) (2010-12). TRE was indicated by 4-12 hours of daily eating window (EW). The adherence to TRE was assessed by summing up the TRE days in a week. We also calculated 7-day averaged values of EW hours, last/first time of eating occasion (EO), and within-person variability of these variables. Weight change from baseline to a follow-up survey (NHS: 2016-18; NHSII: 2015-17) were analyzed. Total energy expenditure (TEE) and physical activity expenditure (PAEE) were measured at baseline using the doubly labeled water dilution method. Results: Longer averaged hours of EW were related to greater weight gain (β 0.7 [0.3] kg per 2 hours) after adjusting for covariates of demographic factors, total energy intake, physical activity, alcohol, within-person variation of EW, the averaged time of last EO, and the initial body weight ( p =0.02). Consistently, adherence to TRE was associated with less weight gain (β -0.25 [SE 0.1] kg per day increment; p =0.03) in a model adjusting for these covariates. We found that lower adherence to TRE and longer hours of EW were associated with weight gain ( p <0.05 for both) when the last eating time occurred at a later time in the day than at an earlier time. The baseline energy metabolism (TEE and PAEE) modified the associations between the hours of EW and weight changes ( P interaction-EW-TEE = 0.006; P interaction-EW-PAEE =0.01), showing significant relationships particularly among women with higher levels of adiposity and greater energy expenditure at baseline. Conclusions: Adherence to TRE and fewer habitual hours of eating were related to long-term weight changes among middle-aged and elderly women. The last eating time and energy homeostasis may partly modify the associations.