<p> When existing buildings are reinforced against earthquakes, concrete roughened through chipping (herein simply referred to as “roughened concrete”) and post-installed anchors are often used to join reinforcing parts, such as steel braces and vibration control braces, to the structures. Roughened concrete increases the rigidity of the interlocking resistance and bearing resistance, and it can also increase the shear strength.</p><p> The authors have proposed the use of cylindrical shear-keys in place of roughened concrete as a joining technique. This combination of a cylindrical shear-key and a post-installed anchor is expected to reduce the shear displacement at the joint surface and increase the shear strength. However, when these techniques are used in combination, the stress state of the joint is complex, and the mechanical behavior of the cylindrical shear-keys and post-installed anchors are different. Therefore, it is necessary to verify whether the combination can be evaluated using a mechanical model that adds the mechanical model of each technique (herein termed “pre-modified model”). The purpose of this study is to perform shearing tests with a test specimen composed of a cylindrical shear-key and a post-installed anchor used in combination, to estimate the axial stress state at the joint from the test results, and to propose a mechanical model for the combination (herein termed “modified model”) reflecting these effects.</p><p> The test results showed that when a post-installed anchor and a cylindrical shear-key were used in combination, the value of the aperture displacement δ_V exceeded the value of δ_V when a single post-installed anchor was utilized. Therefore, because the axial force T at the anchor bolt increased when used in combination, we observed a tendency for the bending moment M to decrease. From these test results, the shearing force ^C_δQ_a resisted by the post-installed anchor when used in combination was estimated to be approximately 0.7 times the value when a single post-installed anchor was used. Furthermore, as the average compressive stress _aσ₀ at the joint surface added by the vertical component of T had an average value of −0.60 N/mm², this axial stress should be considered in the mechanical model of the cylindrical shear-key when used in combination. The shearing force transmitted by the cylindrical shear-key when used in combination could have increased owing to these reasons. Furthermore, considering that the mechanical behavior of the cylindrical shear-key was exhibited in the gradually increasing load region until the anchor bolt yielded when used in combination, the shear displacement, δ_max1, was 0.63 mm when it reached the shear strength based on the mechanical model of the cylindrical shear-key. The average value of δ_H when the anchor bolt reached its yield, resulted in a value exceeding δ_max1 = 0.2 mm when a single anchor bolt was used.</p><p> In this study, we performed a detailed examination, focusing on the stress state at the joint when a post-installed anchor and a cylindrical shear-key were used in combination. We have shown that the modified model can be used to predict experimental values at higher accuracy, compared with the pre-modified model.</p>