Effective anticoagulation regimens are needed to reduce risks of thrombosis and bleeding in animal models of ventricular assist device to verify its hemocompatibility, biologic safety and reliability. This study is to develop a validated anticoagulation procedure for a sheep model to test the newly developed CH-VAD. CH-VAD models were established in six healthy sheep by constructing blood bypass of left ventricle -> ventricular assist device -> descending aorta. Heparin infusion was used during operation and in the prior 4 days to maintain activated clotting time 1.5-2.0 times the baseline. From the third day, proper dosage of warfarin was used orally to maintain international normalized ratio values within the range of 1.2-2.0. After termination, we examined whether there was thrombosis in the blood pump, grafts, and anastomotic stoma. Macroscopic and histopathologic examinations were performed in major organs to check for congestion and infarction. Bleeding complications were not found in any animals throughout the experiments. Activated clotting time values were 326 +/- 33 s intraoperatively and 157 +/- 28 s in the prior 4 days postoperatively. Activated partial thromboplastin time values increased slowly and reached the lower limit of the target range on the fourth day. Only in one of six cases was thrombus or fibrosis tissue found in the blood flow channel of the pump. Pathologic analysis showed no thrombosis, necrosis and microembolus in end-stage organs. Under the anticoagulation regimens, coagulation system could be well controlled to avoid thrombosis and bleeding complications in sheep models for CH-VAD.