To investigate the maximum anti-pullout force and stress distribution of cortical bone trajectory (CBT) screws during screw pullout after contacting different cortical bone layers by finite element analysis (FEA) based on Abaqus software, and to provide evidence for increasing screw holding force during CBT screw implantation in clinical practice. Based on the plain CT data of lumbar spine of a healthy male volunteer who visited the Fourth People's Hospital of Guiyang in June 2022, and the standard screw parameters according to cortical bone trajectory screws. A three-dimensional model of L4 vertebral body and CBT screw was established. The diagnostic criteria of osteoporosis by quantitative CT of lumbar spine in China were set as 120mg/cm3 low bone mass model. According to the number of contact layers between screws and cortical bone, the models were divided into group A: CBT screws produced one layer of cortical bone contact with the vertebral body (screw implantation point); group B: CBT pedicle screws produced two layers of cortical bone contact with the vertebral body (screw implantation point + pedicle inner edge); group C: CBT pedicle screws produced three layers of cortical bone contact with the vertebral body (screw implantation point + pedicle inner edge + outer edge of the vertebral body); group D: CBT pedicle screws produced four layers of cortical bone contact with the vertebral body (screw implantation point + pedicle upper wall + pedicle lateral wall + upper edge of the vertebral body); group E: CBT pedicle screws produced five layers of cortical bone contact with the vertebral body (screw implantation point + pedicle posteromedial medial wall + anterolateral wall of the pedicle + upper edge of the vertebral body + outer edge of the vertebral body). According to the reference, after the material assignment was completed, the axial pullout force experiment of screws was simulated on Abaqus CEA engineering software for five groups of finite element models established to observe the maximum axial pullout force of each group of models and the stress distribution of screws and vertebral bodies. For the comparative analysis between each group of models corresponding to the measured data, one-way analysis of variance was used. The stress of cancellous bone failure in finite element model of group E was (8.6 ± 0.9), (8.4 ± 0.9), (8.1 ± 0.9), (8.3 ± 0.8), and (8.8 ± 0.7) MPa, respectively, and there was no significant difference between any group (P > 0.05); the maximum principal stress in the tail of screw was (195.1 ± 35.8), (290.9 ± 32.1), (317.3 ± 44.5), (396.3 ± 51.2), and (526.5 ± 53.1) MPa, respectively, and the stress in cortical bone destruction was (40.6 ± 3.5), (52.6 ± 4.2), (89.4 ± 4.9), (109.0 ± 8.3), and (129.4 ± 6.4) MPa, respectively, and there was significant difference between any group (P < 0.05); the maximum axial pullout force in group A-E was (1890.5 ± 45.0), (1913.4 ± 53.8), (2371.0 ± 108.3), (237.2 ± 43.0), and (119.5 ± 43.0), respectively. The increases were 2%, 16%, 5%, and 7%. There were significant differences between each group and group E (P < 0.05). Axial pullout force increases as the number of contact layers between CBT screws and cortical bone increases. The axial pullout force increases most when the cortical trajectory screw contacts the vertebral body with three layers of cortical bone, and reaches the maximum value when the number of contact layers reaches five layers, regardless of the screw tail stress and the maximum axial pullout force, so it is clinically desirable to make the screw contact more than three layers of cortical bone to obtain higher stability during screw implantation, improve screw stability, and increase the surgical fusion rate.