To evaluate Hyuganatsu oranges (Citrus tamurana Hort. Ex Tanaka) derived arabinogalactan for bone turnover, we performed a randomized placebo-controlled trial. Sixty-three postmenopausal women were age-stratified and randomly assigned to receive arabinogalactan-rich hyuganatsu juice (study group) or a placebo drink (control group) for 90 days. We measured blood tartrate-resistant acid phosphatase 5b (TRACP5b), type I procollagen N-terminal propeptide (P1NP), and other bone turnover biomarker levels at baseline, days 45 and 90 (T90) of the intervention, and day 30 of recovery. Cumulative effects were compared between groups using repeated-measures linear mixed model analysis. The primary endpoint was the difference between the pre- and post-intervention TRACP5b and P1NP levels. Using repeated measures linear mixed model analysis, the study group had significantly lower TRACP5b and P1NP levels at day 90 than the control group (mean [95% confidence interval]; TRACP5b: 310.0 [269.2-350.9] vs. 386.4 [341.2-431.6] mU/dL; P1NP: 53.7 [48.6-58.7] vs. 70.3 [64.1-76.4] ng/mL), whereas other biomarker levels showed no change. Arabinogalactan-rich Hyuganatsu juice suppressed bone mineral turnover and potentially improved ovarian hormone deficiency-induced osteoporosis in postmenopausal women.