For Asian patients with nonvalvular atrial fibrillation (NVAF) using direct oral anticoagulants (DOACs), performance of contemporary various bleeding risk scores in a real-world setting is unknown. The objective of this study was to externally validate the ORBIT bleeding score and the HAS-BLED score in a large pooled real-world Asian population with NVAF using DOACs. We conducted a single-center prospective observational registry of NVAF patients treated with DOACs: the DIRECT registry (UMIN000033283). We assessed predictive and discriminative performance of the ORBIT bleeding and the HAS-BLED scores for major bleeding in 2,216 patients with NVAF using DOACs (63.6% male, median age 73 years, median CHADS₂ score 2). The overall incidence of major bleeding was 4.2% after a median follow-up of 315 days (interquartile range: 76 to 621). The ORBIT bleeding and the HAS-BLED scores both had modest discrimination ability to identify those who had bled versus who had not (C index = 0.64 [95% confidence interval {CI} 0.59, 0.70] and 0.62 [95% CI 0.57, 0.68], respectively). Calibration plots of the ORBIT bleeding score showed similar predictive performance compared with the HAS-BLED score (slope: 0.91 [95% CI 0.40, 1.43] vs 0.72 [95% CI 0.03, 1.40], intercept: 0.24 [95% CI −2.13, 2.61] vs 0.71 [95% CI −2.35, 3.76], respectively). In conclusion, the ORBIT bleeding score and the HAS-BLED score in a real-world of NVAF population with DOACs showed a modest discriminative performance and a similar predictive performance for major bleeding.