Objective: To investigate the safety of warfarin for Kawasaki disease (KD) with coronary artery aneurysm (CAA) and its prognosis. Methods: Twenty one children with KD complicated with giant CAA, multiple CAA in one coronary artery or thrombosis in coronary artery were enrolled in this prospective study. Warfarin was used to control the goal international normalized ratio (INR) ranging from 2.0 to 3.0. The CAA diameter, number, location and thrombus in coronary artery were recorded at the beginning of treatment, 1, 2, 3, 4 weeks and 2, 3, 6, 12 months after treatment, as well as the influence on INR, electrocaroliogram, creatine kinase-MB (CK-MB), troponin I. Standardized warfarin bleeding risk training and management was implemented. Children were divided into implementation group and non-implementation group according to the status of actual implementation of their parents. The incidence of bleeding events was compared between the two groups. Comparisons between groups were performed using a Rank sum test and a Fisher exact test. Results: In the 21 patients (15 males and 6 females), the age of onset ranged from 2 months to 6 years. There were 4 cases with grade Ⅱ, 7 cases with grade Ⅲ, 7 cases with grade Ⅳ and 3 cases with grade Ⅴ according to the severity of coronary arterial lesions before treatment. The time of clinical detection of thrombus in 10 children with thrombosis ranged from the fourth day to the fourth month. The dose distribution of warfarin was 0.06-0.10 mg/(kg·d), and the INR was 1.80-2.59. Among the 10 cases with thrombus, 8 cases had disappearance of thrombi and 2 cases with grade Ⅴ had thrombus organization to different degree. After treatment, the coronary artery ectasia of the 4 cases with grade Ⅱ all returned to normal. Among the 7 cases with grade Ⅲ, 3 cases of coronary artery aneurysms returned to normal, and 4 cases did not change. Among the 7 cases with grade Ⅳ , 5 cases of coronary artery aneurysms shrank to grade Ⅲ, and 2 cases remained unchanged. Three cases with grade Ⅴ lesions had no changes in aneurysm. Neither new thrombus nor new CAA was detected during the treatment. There was no significant change in electrocardiogram before and after treatment. No statistically significant difference was found regarding the troponin I (0.07 (0-3.01) vs. 0.04 (0-0.29) μg/L, Z=0.932, P>0.05) and CK-MB (20.6 (11.2-58.2) vs. 29.0 (16.7-47.0) U/L, Z=1.906, P>0.05) before and after treatment. The incidence of bleeding events in the implementation group was significantly lower than that in the non-implementation group (2/15 vs. 4/6, Fisher=5.689, P=0.031). Conclusions: The application of goal INR of 2.0-3.0 and adjustment of warfarin dose according to the severity of CAA combined with standardized and strict warfarin bleeding risk training and management, can increase the safety of warfarin therapy in children with KD, improve the prognosis of coronary artery lesions, promote the dissolution of thrombi, prevent new thrombosis, and effectively reduce the incidence of bleeding complication.