Objective: To compare three surgical knots for preventing leakage from the vascular bundle during ligation in simulated equine open castrations. Study design: Randomized, case-control, in vitro study. Sample population: Testes (N = 60) collected from 30 horses. Methods: Testes were collected from 30 horses and randomly assigned to one of three groups: group G (friction, giant knot), group T (modified transfixing knot), or group S (sliding, strangle knot; n = 20/group). The assigned knot was used to ligate the vascular bundle during open castration. The length of suture material used and the leak pressure of the testicular artery were measured and compared between groups. Results: Strangle knots consistently leaked at higher pressures (median, 735.5 mm Hg; interquartile range [IQR], 735.5-735.5) compared with giant (median, 441.3 mm Hg; IQR, 367.8-643.6) and transfixing (median, 419.2 mm Hg; IQR, 323.6-643.6; P < .0001) knots. Both the strangle (median, 5 cm; IQR, 4.5-5.5) and giant (median, 6 cm; IQR, 5.35-6.075) knots required less suture material compared with the transfixing (median, 9.2 cm; IQR, 8.425-10.38; P < .0001) knot. Conclusion: The three surgical knots tested withstood pressure well above physiological levels in simulated open castrations. The strangle knot withstood higher pressure and required similar (giant) or less (transfixing) suture material than the other two knots. Clinical significance: This study provides evidence to support the use of a strangle knot to ligate the vascular bundle during simulated open castrations in horses.