To compare the results of thrombelastography (TEG) and the conventional coagulability test in patients with sepsis, and to discuss the value of TEG in monitoring blood coagulation dysfunction in patients with sepsis. The clinical data of 92 adult patients with sepsis admitted to Department of Critical Care Medicine of the First Affiliated Hospital of Guangxi Medical University were retrospectively analyzed. The patients were divided into sequential organ failure assessment (SOFA) score ≥ 12 group (n = 47) and SOFA < 12 group (n = 45). Thirty-five non-sepsis adult patients with normal coagulation function served as control group. The venous blood was collected for conventional blood coagulation test and routine examination of blood, D-dimer, procalcitonin (PCT), and TEG, and the differences were compared among three groups. Correlations between SOFA and various indexes of patients with sepsis were analyzed by Spearman rank correlation method. As shown in the results of the conventional blood coagulation test, D-dimer was gradually increased with the aggravation of the disease, the values in non-sepsis, SOFA < 12, and SOFA ≥ 12 groups were 0.523 (0.273, 0.928), 0.863 (0.673, 4.221), and 4.118 (2.420, 5.653) mg/L respectively (Z = 25.163, P = 0.000). Platelet count (PLT) in SOFA ≥ 12 group was significantly lower than that of the SOFA < 12 group and non-sepsis group [×10(9)/L: 28.6 (12.8, 48.9) vs. 257.3 (152.6, 339.8), 182.0 (118.0, 229.0), both P < 0.01]. There was no significant difference in prothrombin time (PT) and international normalized ratio (INR) among three groups, and it indicated that the conventional blood coagulation test might not respond quickly to the change in coagulation status of sepsis patients. As shown in the results of TEG, the values of reaction time (R value) and kinetics time (K value) in SOFA < 12 group were lower than those of the non-sepsis group [R value (minutes): 4.4 (3.6, 6.1) vs. 6.3 (6.0, 6.7), P < 0.01; K value (minutes): 1.1 (1.0, 1.5) vs. 1.5 (1.3, 1.8), P < 0.05], while they were higher in SOFA ≥ 12 group than those of the non-sepsis group [R value (minutes): 7.0 (5.7, 8.7) vs. 6.3 (6.0, 6.7), P > 0.05; K value (minutes): 4.2 (3.4, 7.1) vs. 1.5 (1.3, 1.8), P < 0.01]. The α angle, maximum amplitude (MA) and coagulation index (CI) in SOFA < 12 group were higher than those of the non-sepsis group [α angle (degree angle): 73.3 (68.5, 74.7) vs. 66.8 (62.2, 69.0), P < 0.01; MA (mm): 71.7 (61.9, 73.3) vs. 60.3 (58.2, 63.8), P < 0.01; CI: 3.1 (-0.1, 3.9) vs. 0.9 (-0.4, 1.3), P < 0.05], while they were lower in SOFA ≥ 12 group than those of the non-sepsis group [α angle (degree angle): 48.1 (36.6, 53.0) vs. 66.8 (62.2, 69.0), P < 0.01; MA (mm): 37.8 (30.0, 45.7) vs. 60.3 (58.2, 63.8), P < 0.01; CI: -5.6 (-8.4, -3.6) vs. 0.9 (-0.4, 1.3), P < 0.01]. The above results indicated that TEG could distinguish quickly the hypercoagulability and hypocoagulability status in septic patients. PCT in non-sepsis, SOFA < 12, and SOFA ≥ 12 groups were 0.27 (0.05, 1.80), 0.68 (0.10, 10.00), 41.10 (4.24, 100.00) μg/L respectively (Z = 195.475, P = 0.000), which indicate the severity of infectious disease. Correlation analysis results showed that SOFA score was negatively correlated with PLT, α angle, MA, and CI (r value was -0.853, -0.833, -0.881, and -0.859, respectively, all P = 0.000), and it was positively correlated with activated partial thromboplastin time (APTT), D-dimer, R value, K value, and PCT (r value was 0.381, 0.561, 0.587, 0.831, 0.775, respectively, P < 0.05 or P < 0.01), and non correlations was founded with PT, fibrinogen (FBG), and INR (r1 = 0.211, P1 = 0.233; r2 = -0.252, P2 = 0.142; r3 = 0.248, P3 = 0.148). TEG can effectively monitor the change in coagulation in patients with sepsis, and distinguish the hypercoagulable and hypocoagulable state. TEG may be a valuable tool to evaluate degree and risk of sepsis objectively.