A real-world association between direct oral anticoagulant (DOAC) concentration and clinical outcomes among Asian patients with atrial fibrillation (AF) is reported herein. Patients with AF aged ≥ 20 years who used DOAC for ≥ 3 days were enrolled. Trough and peak DOAC concentrations were measured and compared with the expected range reported in clinical trials. The Cox proportional hazard model was used to investigate the association between concentration and outcomes. From January 2016 to July 2022, a total of 859 patients were enrolled. Among them, 22.5%, 24.7%, 36.4%, and 16.4% were on dabigatran, rivaroxaban, apixaban, and edoxaban, respectively. Compared with clinical trials, the proportion of DOAC concentrations higher or lower than the expected range were 9.0% and 14.6% for trough, respectively, and 20.9% and 12.1% for peak, respectively. The average follow-up duration was 2.4 ± 1.6 years. The incidence of stroke and systemic thromboembolism (SSE) was 1.31 per 100-person years, and low trough concentration predicted SSE (hazard ratio (HR) = 2.78 (1.20, 6.46)). The incidence of major bleeding was 1.64 per 100-person years, and high trough was associated with major bleeding (HR = 2.63 (1.09, 6.39)). The association between peak concentration and SSE or major bleeding was nonsignificant. Off-label underdosing (odds ratio (OR) = 2.69 (1.70, 4.26)), once daily DOAC dosing (OR = 3.22 (2.07, 5.01)), and high creatinine clearance (OR = 1.02 (1.01, 1.03)) caused low trough concentration. Contrarily, congestive heart failure was significantly associated with high trough concentration (OR = 1.71 (1.01, 2.92)). In conclusion, trough DOAC concentration measurements should be considered among patients at risk of out-of-expected range DOAC concentrations.