The soil–anchor interface mechanical behavior is of practical importance for the bearing performance of ground anchors. Element-scale pullout test is a simple, reliable, and economical technique to investigate the interface shear stress–displacement response of anchors. However, the influence range of boundary effect for the anchoring element-scale pullout specimens is poorly understood, particularly under the actual stress and construction conditions. In addition, there is still a lack of study on the mechanical behavior of soil–anchor interface subjected to ground pressure via element-scale pullout test. In this paper, an element-scale pullout test setup that can simulate ground pressure and construction method of ground anchors was specially developed, with details introduced carefully. A response surface function-based soil dry density prediction model was proposed for the test design. Comparison with the soil consolidation test results indicates that this model equipped with high fitting and prediction accuracy. Twenty groups of element-scale pullout tests with different ground pressures and anchor diameters were conducted. Based on the test design of the proposed model, the soil density and moisture content were almost the same for all specimens after consolidation. The shrinkage of anchor hole, interface shear stress–displacement response, and interface shear strength were discussed adequately. It is shown that the soil–anchor interface shear strength is hardly dependent on ground pressure under the actual construction method of anchors. The influence of boundary effect for the element-scale pullout test can almost be eliminated when diameter ratio of specimen to anchor exceeds 5.0 ~ 6.7.