OBJECTIVE: To examine associations between protein intake per day and at different meals and skeletal muscle mass declines. DESIGN: Two-year prospective cohort study among older community dwellers. SETTING: National Institute for Longevity Sciences-Longitudinal Study of Aging (NILS-LSA) in Japan. PARTICIPANTS: Older men (n 292) and women (n 363) aged 60-87 years who participated in the baseline (2006-2008) and follow-up studies (2008-2010) of NILS-LSA and did not exhibit low skeletal muscle mass at baseline. Muscle mass was assessed using dual-energy X-ray absorptiometry at baseline and follow-up. Low muscle mass was defined as skeletal muscle mass index <7·0 kg/m2 for men and <5·4 kg/m2 for women at follow-up. Daily protein intake and protein intake at each meal were calculated from 3 d dietary records at baseline and sex-stratified tertiles were determined. RESULTS: Mean (sd) protein intake at breakfast, lunch and dinner was 22·7 (7·8), 26·7 (9·3) and 37·4 (10·5) g for men and 19·3 (6·3), 23·2 (7·3) and 28·5 (7·0) g for women, respectively. After adjusting for age, baseline skeletal muscle mass and other confounders in logistic modelling, greater total protein intake was associated with lower prevalence of skeletal muscle mass decline among men at follow-up (P = 0·024). Particularly, the OR (95 % CI) for high lunchtime protein intake was low (0·11 (0·02, 0·61); P = 0·01). No significant association between total protein intake and prevalence of skeletal muscle mass decline was found among women. CONCLUSIONS: High total protein intake, particularly at lunchtime, is associated with retention of skeletal muscle mass in men.