This paper describes a technique for modeling the installation and pullout of a dynamically installed anchor (DIA) using specifically designed equipment on a beam centrifuge. The equipment consists of three subassemblies, an X-Y table, an anchor installation system, and a pullout system. The X-Y table allows the location of anchor installation and pullout direction and pullout angle to be adjusted within a relatively short time without stopping the centrifuge. This permits the pullout angle with respect to vertical and the azimuthal angle of the pullout direction relative to the fins to be adjusted under high-g conditions. Inclined anchor penetration could also be preset by changing the lateral offset of the top of the guide tube. Finally, a very short transition time of several seconds between installation and pullout can be achieved, thereby allowing nearly short-term conditions to be achieved during pullout. The centrifuge model test results show that vertical pullout of non-vertically installed anchors and non-vertical pullout of vertically installed anchors both give a higher holding capacity than vertical pullout of vertically installed anchors. This indicates that the transverse pullout resistance of an anchor may contribute measurably to the anchor capacity and raises the possibility that, when appropriately installed, even higher pullout capacity may be obtainable from inclined pullout of non-vertically installed anchors. The findings also suggest that the ratio of short-term to long-term capacity may differ for different soils, thereby underlining the importance of being able to conduct tests at sufficiently shorter reconsolidation times.