The seismic response of historical masonry buildings is largely controlled by the effectiveness of wall-todiaphragm connections, which can be improved using injection anchors. Despite their use, there is a need for a better understanding of the performance of such anchoring system in stone masonry walls. This paper presents quasi-static pull-out tests performed on twelve specimens to investigate the behaviour of injection anchors in rubble stone masonry walls when breakout failure occurs. For each specimen, the experimental results are presented in terms of force-displacement curves, propagation of damage and crack pattern. It is shown that the anchoring details adopted in this study had a negligible influence on the pull-out force capacity of the anchoring system, while an increase in peak pull-out force was observed with increasing overburden stress applied. Because stone masonry specific capacity formulations have not yet been presented in the literature, the capacity is predicted using state-of-the art formulations for the pull-out load capacity of anchors installed in brick masonry. The limits of their applicability are discussed using the obtained experimental results.