Objective: In pedicle screw fixation,the heads of monoaxial screws need to be directed in the same straight line to accommodate the rod placement by backing out during operation, which decreases the insertional torque and internal fixation strength. While polyaxial screws facilitate the assembly of the connecting rod, but its ball-in-cup locking mechanism reduces the static compressive bending yield strength as compared with monoaxial screws. Our study aimed to assess the mechanical performance of a modified pedicle screw.Methods: In this study, the tail of the screw body of the modified pedicle screw was designed to be a cylindershaped structure that well matched the inner wall of the screw head and the screw head only rotated around the cyclinder. Monoaxial screws, modified screws and polyaxial screws were respectively assembled into 3 groups ofvertebrectomy models simulated by ultra high molecular weight polyethylene (UHMWPE) blocks. This model was developed according to a standard for destructive mechanical testing published by the American Society for Testing Materials (ASTM F1717-04). Each screw design had 6 subgroups, including 3 for static tension, load compression and torsion tests, and the rest for dynamic compression tests. In dynamic tests, the cyclic loads were 25%, 50%, and 75% of the compressive bending ultimate loads respectively.Yield load, yield ultimate load, yield stiffness, torsional stiffness, cycles to failure and modes of failure for the 3 types of screws were recorded. The results of modified screws were compared with those ofmonoaxial and polyaxial screws.Results: In static tests, results of bending stiffness,yield load, yield torque and torsional stiffness indicated no significant differences between the modified and monoaxial screws (P>0.05), but both differed significantly from those ofpolyaxial screws (P