Abstract: Objective To screen for the optimal dose of benzene and cyclophosphamide using an orthogonal design for establishment of New Zealand rabbit models of aplastic anemia. Methods Following an orthogonal experimental design, the effects of 3 levels of 4 factors, namely the dose of benzene (A), the dose of cyclophosphamide (B), the number of benzene injections (C), and the number of cyclophosphamide injections (D) were tested in the establishment of New Zealand rabbit models of aplastic anemia using a L9 (34) orthogonal table, and the optimal protocol for the model establishment was selected from the 9 experimental groups. Each rabbit received subcutaneous injection of benzene on the back every other day, followed by daily cyclophosphamide injection via the ear vein for prescribed times. The blood routine was examined every 6 days, and before modeling and at 36 days after modeling, a small sample of the femoral bone was collected for bone marrow histopathological examination. Results Comparison of the white blood cell, erythrocyte and platelet counts among the 9 groups showed successful modeling in Groups 4-9, and daily mean reduction rates of the cell counts in Groups 7, 8, and 9 differed significantly from those in the other groups (P<0.05). In Group 7, bone marrow sections showed low myelodysplasia, reduced hematopoietic tissue, reduced or even absence of megakaryocytes, and increased fat cells. Further observation found that the rabbits in Group 7 had sustained bone marrow suppression, consistent with the clinical characteristics of the disease. Conclusion Stable models of aplastic anemia can be established efficiently in New Zealand rabbits by a combination of 8 subcutaneous injections of benzene at 1.5 mL/kg and 4 intravenous injections of cyclophosphamide at 10 mg/kg.