OBJECTIVE: To compare the biomechanical characteristics of thoracolumbar fractures treated with uniplanar screws, monoaxial screws, and polyaxial screws using finite element analysis. METHODS: CT data of the thoracolumbar spine T-L from a healthy volunteer were collected, and using finite element software, models of both normal and fractured spines were created. Three different fixation models were constructed with monoaxial screws (Mps group), polyaxial screws (Pps group), and uniplanar screws (Ups group), respectively. The L vertebra was fixed and a compressive load of 150 N and a torque of 10 N•m were applied at the T end to simulate flexion, extension, lateral bending, and rotation movements of the spine. The range of motion (ROM) and internal fixation stress within the screws and rods were recorded for each movement direction. RESULTS: A finite element model of the healthy human spine T-L was established and validated for accuracy. All three fixation models demonstrated decreased ROM in all tested movements. The UPS group showed the lowest percentage of ROM in flexion, extension, and lateral bending movements, with a mid-range percentage of ROM in rotation, and relatively the best stability. The PPS group had the highest ROM percentages in all directions of movement, with the worst relative stability. The maximum von Mises stress for pedicle screws and rods in all fixation modes occurred in flexion, with the MPS group's screws showing significantly higher stress peaks in flexion and both rotations than those of the PPS and UPS groups. The rods of the PPS group had significantly lower stress peaks in all motion states compared to those of the MPS and UPS groups. CONCLUSIONS: Uniplanar screws can effectively distribute stress, reduce the risk of screw and rod breakage, and ensure stability of spinal fixation.