Objective: To evaluate the therapeutic effect of the fludarabine and cytarabine (FA) regimen on acute myeloid leukemia (AML) at different phases during treatment. Methods: A total of 185 patients with AML were divided into 4 groups based on the outcome of previous treatments. Patients in Group 1 had no remission after the first course of induction chemotherapy (n = 55). Patients in Group 2 had no remission after no less than two courses of induction chemotherapy (n = 41). Patients in Group 3 had early relapse (n = 40). Patients in Group 4 had late relapse (n = 49). Patients in groups 2, 3 and 4 had refractory AML or AML with relapse. We assessed the efficacy and toxicity of FA combination chemotherapy in each of these 4 groups. Results: The complete remission (CR) rates of Groups 1, 2, 3 and 4 were 74.5% (41/55), 45.9% (19/41), 17.5% (7/40) and 38.8% (19/49), respectively. The CR rate was higher in Group 1 than in the other 3 groups (34.6%, 45/130) (P = 0.000). A significant correlation was found between CR rate and the number of chemotherapeutic courses (P = 0.023). The main adverse reactions included bone marrow suppression and secondary infection. Conclusion: FA regimen is a good choice for patients with AML, especially those who have failed to achieve CR after the first course of induction chemotherapy. Objective: To evaluate the therapeutic effect of the fludarabine and cytarabine (FA) regimen on acute myeloid leukemia (AML) at different phases during treatment. Methods: A total of 185 patients with AML were divided into 4 groups based on the outcome of previous treatments Patients in Group 1 had no remission after the first course of induction chemotherapy (n = 55). Patients in Group 2 had no remission after no less than two courses of induction chemotherapy (n = 41). Patients in Group 3 had early relapse (n = 40). Patients in groups 2, 3 and 4 had refractory AML or AML with relapse. We assessed the efficacy and toxicity of FA combination chemotherapy in each of these 4 Results: The complete remission (CR) rates for Groups 1, 2, 3 and 4 were 74.5% (41/55), 45.9% (19/41), 17.5% (7/40) and 38.8% 49), respectively. The CR rate was higher in Group 1 than in the other 3 groups (34.6%, 45/130) (P = 0.000). A significant correlatio n was found between CR rate and the number of chemotherapeutic courses (P = 0.023). The main adverse reactions included bone marrow suppression and secondary infection. Conclusion: FA regimen is a good choice for patients with AML, especially those who have failed to achieve CR after the first course of induction chemotherapy.