A large amount of coal gangue is generated during coal excavation and preparation. The coal gangue accumulates on the ground surface and is easily affected by environment and causes disasters. The mechanism of the effects of drying-wetting cycles and stone content on the strength characteristics of coal gangue is investigated through large-scale direct shear tests. The influence of proportion of rock( ) and the number of drying and wetting on the volume and relative breakage rate( ) of soil-rock mixture was revealed. The equation of shear strength degradation of soil-rock mixture under drying-wetting cycles was established. The test results showed that as the rock content increases，the internal friction angle of the coal gangue soil-rock mixture increased linearly，while the cohesion exhibited a downward trend. The gangue soil-rock mixture exhibited obvious shear shrinkage during the shear process. The shear strength and maximum normal strain of the mixture increased with an increase in rock content but decrease with the number of drying-wetting cycles. The shear strength of the soil-rock mixture significantly deteriorated during the first drying-wetting cycle，and it did not change significantly after three drying-wetting cycles. The drying-wetting cycle enhanced the fragmentation tendency of rock in the soil-rock mixture，raised the  of the rock，and increased the content of soil particles(particle size＜5 mm). The   increased with the number of drying-wetting cycles，while under low normal stress( = 200 kPa)，the drying-wetting cycles had a limited impact on rock fragmentation.

Based on the Duncan-Chang model and the equivalent additional stress effect of geocell reinforcement，an equivalent strength and stiffness model tailored for shear-contractive soils reinforced with geocells has been formulated. This model was further developed using the UMAT subroutine module of the ABAQUS software. To validate the effectiveness of the model and the correctness of the UMAT subroutine，triaxial tests and foundation bearing capacity model tests were conducted. The results reveal that at low axial strains，the strength ratio   between reinforced and unreinforced soils is less influenced by natural density. However，as the axial strain increases，clays with higher natural densities will exhibit a greater   value. The developed UMAT subroutine shows good agreement with experimental data up to an axial strain of 0.15，with minor deviations thereafter but overall low relative errors，confirming the validity of the reinforced clay model and the correctness of the secondary development of the UMAT subroutine. Prior to reaching peak strength，both the proposed model and the discrete reinforcement-soil model align well with experimental results. After reaching peak strength.，the numerical simulation results of the proposed model are closer to experimental outcomes. Additionally，this model offers advantages of simplified modeling and high computational efficiency. For settlement values S≤4 mm，the load-displacement curves of both unreinforced and reinforced sandy soil models align well with experimental data. When S＞4 mm，the calculated bearing capacity of the models is slightly higher than experimental data. The research outcomes of this study can provide a model reference for geocell-reinforced shear-contractive soils and offer new methods for finite element analysis in their engineering applications.