Tail prehension is a common, although poorly studied behavior among seahorses. We investigate this behavior in the potbellied seahorse, Hippocampus abdominalis, and the longsnout seahorse, H. reidi. Hippocampus abdominalis has a slender tail and significantly higher number of tail segments (45-48) than H. reidi (33-37). We hypothesize that the tail of H. abdominalis would be more flexible than the shorter tail of H. reidi. We compared 3D grasping kinematics on a 1cm horizontal perch. In H. reidi the whole tail is involved in grasping with an increased range of motion towards the tip. In contrast, in H. abdominalis the most proximal third of the tail is not involved in grasping. Still, other kinematic variables are similar for the two species. In addition, both species show lateral bending during tail curling, an unexpected finding that might be important for modulation of grasping in substrates with different orientations, such as corals and seagrasses. Different artificial holdfasts were also tested in a preference study in H. abdominalis. Seahorses selected for vertical oriented, cylindrical and smooth holdfasts. However, color (sand vs. green) and holdfast diameter (1 and 1.5cm) were neutrally selected for. Preference for vertical holdfasts is likely a result of the relative abundance of similarly oriented substrates in the wild, and thus selection for lateral bending may have played a role in prehensile tail evolution in seahorses. Pot-bellied seahorses also selected negatively for rough and blade like structures, which is likely explained by increased contact area in smooth, cylindrical surfaces which should facilitate attachment. Negative selection for rough structures seems to indicate that friction mechanisms are not predominant, while selection for higher contact area suggests reliance on wet adhesion and muscular grasping.