Objective: To investigate the distribution of bone marrow lymphocyte subsets in patients with acute myeloid leukemia (AML),and comparing the bone marrow immune function between patients at different stage of disease and with different prognostic risk, and exploring their related clinical significance. Methods: The bone marrow samples from 131 patients with AML before treatment were collected, and the lymphocyte subsets CD4⁺ T cells, CD8⁺ T cells and CD19⁺ B cells were identified by multi-color flow cytometry for determining the percentages of lymphocyte subsets and CD4⁺/CD8⁺ ratio. The results were compared with that of healthy bone marrow samples obtained from the public data platforms. The AML patients included de novo cases (94 cases), secondary cases （18 cases）, and relapsing cases after chemotherapy （19 cases）. The de novo AML were categorized into low risk, intermediate risk and high risk groups according to the cytogenetic abnormality. Differences in lymphocyte subsets between the three groups and between the three different risk levels were compared. In relapsing patients, the percentages of lymphocyte subsets at onset were compared with that at relapse. Results: Compared with healthy subjects, AML patients had a higher percentage of CD8⁺ T cells (31.73%±12.38% vs 21.40±7.33%, P<0.001) and lower percentage of CD19⁺ B cells ( 9.62% vs 14.03%, P<0.01) and lower ratio of CD4⁺/CD8⁺ (1.04 vs 1.48, P<0.05). When compared with de novo patients, relapsing patients had higher CD8⁺ T cells (41.56%±11.64% vs 29.86%±12.20%, P<0.001), but lower CD19⁺ B cells ( 4.18% vs 11.82%, P<0.05) and CD4⁺/CD8⁺ ratio (0.59 vs 1.12, P<0.05). Paired comparison showed that relapsing patients had significant lower percentage of CD19⁺ B cells than that at onset(2.40% vs 12.41%, P<0.05). There were no significant differences in distribution of lymphocyte subsets between different risk level groups of de novo AML. The percentages of NK cells between the three groups were not different significantly. Conclusions: The bone marrow of AML patients showes obvious suppression of humoral immunity and cellular immunity. The suppression of bone marrow immune function is obviously more severe in patients with relapsing AML than in de novo. It is suggested that the distribution of bone marrow lymphatic subsets may be an effective indicator for detecting and assessing the prognosis of the disease.