OBJECTIVE: The aim of this study is to investigate the biomechanical changes in the sandwich vertebrae (SV), fractured vertebrae, and adjacent vertebrae at the thoracolumbar vertebrae in patients with osteoporotic vertebral compression fracture (OVCF) who underwent several percutaneous vertebroplasties (PVP) with varied cement volumes. METHODS: The finite element (FE) model of the T10-L2 thoracolumbar vertebral body is established. The augmented vertebrae (AV) of T11 and L1 is simulated and cylindrical bone cement is placed vertically in its center. The models are categorized into four types according to the volume of bone cement, 2mL bone cement group (model A), 4-mL bone cement group (model B), 6-mL bone cement group (model C), and 8-mL bone cement group (model D). By applying 500 N axial load on the upper surface of T10 and fixing the lower surface of L2, the maximum von Mises stress of the vertebrae and the maximum displacement of the sandwich vertebrae are analyzed and compared. RESULTS: The maximum von Mises stresses of the T11 and L1 augmented vertebrae of Model C are lower than those of the fractured vertebrae of Models A and B in all directions of activity. The von Mises stresses of the augmented vertebrae of Model C and Model D are similar. The von Mises stresses of the fractured adjacent vertebrae T10 and L2, and the sandwich vertebrae T12 do not change significantly with the change in cement volume. In addition, the von Mises stress of T12 is lower than that of T10 in all four groups. The minimum value of T12 displacement in Model C is 3.0 mm. CONCLUSION: Under the condition of no leakage, the stress distribution of the AV can be optimized by expanding the supporting area of bone cement to about 6 ml, which not only reduces the risk of recurrent fractures of adjacent vertebrae and AV, but also prolongs the service life of the implants by reducing the stress of bone cement, which provides the basis for the appropriate amount of bone cement required for clinical multi-level PVP.