A micromechanics computational algorithm for fibrous composites including fiber, matrix and interface is introduced to study the impact of interface on composite behavior. The domains are modeled by finite elements with the interface simulated by cohesive zone elements. The constitutive of the cohesive zone behavior is extracted from the experimental traction-separation relations. By implementing this model under different loading conditions, significant difference in the composite behavior is observed with different cohesive zone laws and different fiber waviness. The composite strength and stiffness for the examined cases are compared to experimental data and are in good agreement. The procedure is then extended to study the impact of adhesion on brain axonal injury. The constituents of the brain tissue are modeled as linear viscoelastic materials. This micromechanical model has been implemented to study the impact of adhesion and waviness on the stiffness and viscous behavior of brain tissue with respect to time.