This study was undertaken to compare the effects of acute hemodilution (AH) induced separately by three solutions on blood properties and hemodynamics. Sixty-six patients, ASA grade Ⅰ to Ⅱ, aged 18 to 54 years, having open heart surgery with fentanyl-pancuronium anesthesia, were acted as subjects. During the induction of AH, blood 10-12ml?kg~(-1) was drawn from internal jugular vein in each subject, as Ringer's lactated injection (RL) at equivalent volume was infused intravenously, afterwards, all subjects were randomly allocated to being intravenously infused with RL at a dose of two times as many as the blood volume drawn (BVD) (group RL), Haemaccel 35 (R35) (group H35) or hydroxyethyl starch (Hes) (group Hes) at a dose equal to BVD, respectively. All patients were observed for 30 mins following the status of AH. The results showed that plasma colloid osmotic pressure was descended markedly in these three groups, but more significantly in group RL immediately after induction of AH, and was kept at low level in group H35 and RL and continued to decrease in group Hes during AH. The hemodynamic values of these three groups were within normal range during whole procedures. Prothrombin time and activated partial thrombin time were prolonged in each group, but within normal range. No anaphylactic reactions occured in all subjects. It is suggested that H35 can be applied more safe to acute moderate hemodilution in clinical anesthesia.