Introduction: Osteosarcoma is the most common primary malignant bone tumor, and the grading of osteosarcoma cells relies on traditional histopathology and molecular biology methods, which require RNA extraction, protein isolation and immunohistological staining. All these methods require cell isolation, lysis or fixation, which is time-consuming and requires certain amount of tumor specimen. In this study, we report the use of Raman spectroscopy for grading of malignant osteosarcoma cells. Methods: We demonstrate that, based on the detection of differential production of mineral species, Raman spectroscopy can be used as a live cell analyzer to accurately assess the grades of osteosarcoma cells by evaluating their mineralization levels. Mineralization level was assessed by measuring amount of hydroxyapatite (HA), which is highly expressed in mature osteoblasts, but not in poorly differentiated osteosarcoma cell or mesenchymal stem cells, the putative cell-of-origin of osteosarcoma. Results: We found that under Raman spectroscopy, the level of HA production was high in MG-63 cells, which are low-grade. Moreover, hydroxyapatite production was low in high-grade osteosarcoma cells such as 143B and SaOS2 cells (p < 0.05). Matrix metalloproteinase MMP2, MMP9 were highly expressed in SaOS2, 143B and MSCs and decreased in human fetal osteoblast (FOB) and MG-63 cells as expected (p < 0.05). These results may highlight the inverse correlation between HA level and prognosis of osteosarcoma. Conclusions: The use of Raman spectroscopy for the measurement of HA production by the protocol reported in this study may serve as a useful tool to rapidly and accurately assess the degree of malignancy in osteosarcoma cells in a label-free manner. Such application may shorten the period of pathological diagnosis and may benefit patients who are inflicted with osteosarcoma.