Abstract: In order to calibrate the discrete element parameters of viscoplastic soil based on the Edinburgh Elasto-Plastic Adhesion (EEPA) model, two constants to describe the stress-strain behavior of the sample soils were defined in this study. The axial strain of the soil under a load of 300 kPa was used to characterize the plasticity of the soil in the uniaxial closed compression test. First, two EEPA model parameters based on the physical properties of the soil and other scholars' research were determined, i.e. the constant pull-off force and the load branch index. Then, another two factors, surface energy and plastic deformation ratio, that had significant effects on axial strain and unconfined compressive strength based on the Plackett-Burman test results were described. Next, the central composite test based on the response surface method was designed, and the two factors that matched the actual measured axial strain and unconfined compressive strength based on the test results were determined. According to the test results. the four discrete element parameters corresponding to the two measured constants, unconfined compressive strength and axial strain were solved. Finally, the soil discrete element parameters were calibrated based on the EEPA model, and those were that plastic deformation ratio of 0.36, surface energy of 15.6 J/m2, static friction coefficient of 0.6, rolling friction coefficient of 0.26, recovery coefficient of 0.37, adhesion branch index of 4.24 and tangential stiffness factor of 0.52. Verification test results showed that the EEPA model parameters calibrated based on the response surface method could simulate the plastic deformation of the sample soil under a load of 300 kPa and the stress-strain behavior within 3%-45% of the axial strain. In addition, the results of Quadratic orthogonal rotation combined test were analyzed and it showed that the tangential stiffness factor was one of the key parameters affecting axial strain and unconfined compressive strength, and the random error of the unconfined compressive strength simulation test was also one of the reasons. Moreover, it was found that the limit of the value range of the tangential stiffness factor was the cause of the huge difference between the simulation and the measured values of the unconfined compressive strength.