<p>Cortical bone fracture may result from various loading conditions, such as tension, compression, and bending. Different loads would alter the ratio of normal-to-shear strain in osteons, consequently influencing the failure criterion for mechanical state assessment. While the choice of an unsuitable failure criterion may lead to reduced prediction fidelity, few studies have systematically examined the differences arising from various failure criteria in cortical bone fracture simulations. To address this gap, this study investigated the effects of four failure criteria on cortical bone fracture simulations under three-point bending and compression. The results indicate that the principal and invariant strain failure criteria provided accurate fracture predictions under compression, and the equivalent strain and invariant strain criteria yielded reliable results under bending. Prediction accuracy depends fundamentally on the congruence between the strain growth rate in the criterion and the actual deformation behavior of osteons in specific loading environments. This growth rate varies with loading conditions due to differences in strain calculation formulations. Consequently, the optimal failure criteria differ between compression and bending. This study demonstrates the effectiveness of cortical bone fracture simulation across diverse loading conditions and elucidates numerical simulation variations attributable to different failure criteria. These findings offer valuable insights for enhancing cortical bone fracture prediction accuracy, providing a foundation for future research and potential clinical applications.</p>