Low protein intake can lead to declined lean mass (LM) in elderly. We examined the associations of total protein (TP), animal protein (AP) and plant protein (PP) intakes with LM. The association of TP intake with LM change was further evaluated according to weight change status. This cross-sectional and prospective cohort study included 554 women aged 68 (sd 1·9) years from the Osteoporosis Risk Factor and Prevention – Fracture Prevention Study (OSTPRE-FPS). The intervention group (n 270) received daily cholecalciferol (800 IU; 20 μg) and Ca (1000 mg) for 3 years while the control group received neither supplementation nor placebo (n 282). Participants filled out a questionnaire on lifestyle factors and a 3-d food record in 2002 and underwent dual-energy X-ray absorptiometry for body composition measurements at baseline and 3 years. Multiple linear regressions evaluated the association between protein intake and LM, adjusting for relevant covariates. At the baseline TP and AP intakes were positively associated with LM and trunk LM, TP was associated also with appendicular LM (aLM). Follow-up results showed that in the total population and the intervention group, higher TP and AP were associated with increased LM and aLM (P ≤ 0·050). No such associations were observed in the control group. PP intake was also associated with aLM change in the total population. Overall, the associations were independent of fat mass. Further, among weight maintainers, TP intake was positively associated with LM, aLM and trunk LM changes (P ≤ 0·020). In conclusion, dietary TP, especially AP, intake may be a modifiable risk factor for sarcopenia by preserving LM in the elderly.