Abstract: Geogrid has been extensively used in reinforced soil engineering such as embankments, slopes, retaining walls. Analysis on the interfacial mechanical interaction between the reinforcement and the soil is a key factor for understanding action mechanisms. This paper aims to investigate the actual shear stress-displacement relationship of the pull-out test. The elastic-exponential softening model is established, which considers the progressive failure mechanism and the nonlinear characteristic of the interface according to the stress state of the geogrid subjected to pull-out loads. It is shown that most of the existing calculating models estimate the shear stress of the interfaces. Combined with the governing equation of the interface, the theoretical solutions of the stress state of the geogrid at different stages are deduced. The evolution and distribution law of the interface friction was analyzed in detail. The influence of design parameters on the interfacial shear resistance was investigated. The design parameters include shear stiffness, tensile stiffness, length of reinforcement and exponential attenuation coefficient. The results show that when the interface is in the elastic stage of the pulling process of geogrids, the non-uniformity of interfacial shear stress and the maximum shear stress increases with the increase of shear stiffness, while the modulus of elasticity is opposite. In the softening stage, the longer reinforcement length has more obvious softening of the interface. The shorter reinforcement length may cause the uniform distribution of the interfacial shear stress. The greater exponential attenuation coefficient, the more obvious interfacial shear stress fluctuation. The interfacial shear stress would move to the pulling end of geogrids. After entering the residual stage, the interfacial shear stress increases from the pulling end to the free one and tends to reach the residual stress gradually. For the reinforced soil engineering, the corresponding conclusions as a theoretical basis for understanding reinforcement mechanisms can be applied to the principle of the selection of geogrid.