Background: This study aimed to evaluate the clinical significance of pre-treatment Naples prognostic score (NPS) in patients with osteosarcoma. Methods: The clinical data of 133 osteosarcoma patients between January 2011 and February 2018 in our hospital was retrospectively collected and analyzed. NPS was calculated from four parameters, including serum albumin level, serum total cholesterol (TC), lymphocyte-to-monocyte ratio (LMR), and neutrophil-to-lymphocyte ratio (NLR). Patients were divided into three groups (group 1-3) based on NPS. The relationships between NPS and clinical features, overall survival (OS), and progression-free survival (PFS) were analyzed. Two prediction models based on NPS and clinical parameters were developed: clinical parameters model (model A), and the combined model of NPS and clinical parameters (model B). Their predictive performances were further evaluated and compared. Results: The median follow-up time of this cohort was 46.0 (range, 5-75) months, while the median OS and PFS was 40 (range, 5-75) months and 36 (range, 5-71) months, respectively. NPS was significantly correlated with gender, tumor location, Enneking stage, pathological fracture, local recurrence, and metastasis (all P < 0.05). Variables of NPS, Enneking stage, local recurrence, metastasis, and NLR were confirmed as independent prognostic factors for OS and PFS by univariate and multivariate Cox analysis. Prediction model B obtained larger AUCs for OS and PFS and showed better consistency between nomogram-predicted and actual survival than that of model A at the follow-up time of 1-, 3-, and 5-year. Conclusions: NPS was a novel, reliable, and multidimensional prognostic scoring system with favorable predictive performance for patients with osteosarcoma.