Osteosarcoma is the most common primary malignant bone tumor, and its standard treatment is a combination of surgery and chemotherapy. A poor response to chemotherapy causes unfavorable oncological outcomes. We investigated the correlation between osteoclast differentiation in biopsy specimens and the efficacy of neoadjuvant chemotherapy in resected specimens. Forty-nine patients who underwent neoadjuvant chemotherapy and subsequent surgical treatment at our institution between 1999 and 2018 were enrolled. Using medical records, we investigated the age, sex, tumor size, location, subtype, staging, chemotherapy agents (doxorubicin, cisplatin, ifosfamide, and methotrexate), number of neoadjuvant chemotherapy courses, number of osteoclasts in biopsy specimens, and efficacy of neoadjuvant chemotherapy according to the Rosen and Huvos classification (Grade I-IV) in resected specimens. Univariate and multivariate analyses were performed to identify factors predictive of a good response in resected specimens after neoadjuvant chemotherapy. A good response (Grade III/IV) was detected in 25, while a poor response (Grade I/II) was detected in 24. According to the multivariate analysis, ≥ 46 years old (odds ratio [OR], 0.05; 95% confidence interval [CI], 0.01–0.45; p < 0.01) and ≥ 5 mature osteoclasts in a biopsy specimen (OR, 36.9; 95% CI, 6.03–225; p < 0.01) were significantly associated with the neoadjuvant chemotherapy efficacy. The accuracy for predicting a good response to chemotherapy based on ≥ 5 osteoclasts in a biopsy specimen in patients < 46 years old was 85%. The number of mature osteoclasts in biopsy specimens is a simple factor for predicting the efficacy of chemotherapy before treatment, although further studies will be required to determine the underlying mechanism.