Abstract: An experiment was conducted to compare the effects of two mouse thrombocytopenia models induced by cyclophosphamide at two different administration routes to determine a proper cyclophosphamide administration route that could cause stable thrombocytopenia. A suitable drug dosage that could induce thrombocytopenia in mouse efficiently with the definite administration route was then investigated. BALB/c mice were randomly divided into Normal, Model A and Model B groups. To Model A, 200 mg/kg of cyclophosphamide was given by vena caudalis injection as first dose and 30 mg/kg as maintenance dose by intraperitoneal injection at the following 6 days. To Model B, 150 mg/kg of cyclophosphamide was given by subcutaneous injection once a day for consecutive 3 days. All groups were under investigation for 15 days. The result suggested that a decrease in the number of blood platelets of Model B at the 7th day were significantly than that of Normal. Other platelet related indices like platelet distribution width, mean platelet volume and platelet-large cell ratio of Model B increased significantly in comparison with those of Normal group. The platelets count was reduced but fluctuated greatly, and more than half of the mice died in Model A. Therefore, subcutaneous injection of cyclophosphamide for 3 days was used for the cyclophosphamide dosage test. BALB/c mice were randomly divided into Normal, cyclophosphamide low dose (100 mg/kg), medium dose (120 mg/kg) and high dose (140 mg/kg) groups. All groups were under investigation for 11 days. Though all 3 dosages successfully initiated thrombocytopenia as the platelets number dropped at the 7th day, the low dose was considered to be a suitable one that was of high efficacy and low toxicity. Thus, BALB/c mice challenged by subcutaneous injection of cyclophosphamide 100 mg/kg per day for 3 consecutive day is one simple, feasible and stable mouse thrombocytopenia model that could be used for pharmacodynamic test of the drugs which are supposed to have platelets increasing effect.