Background: The optimal tool for risk prediction of venous thromboembolism (VTE) in inpatients with acute exacerbation of chronic obstructive pulmonary disease (AECOPD) is still unknown. This study aimed to evaluate whether D-dimer could predict the risk of VTE in inpatients with AECOPD compared to the Padua Prediction Score (PPS). Methods: Inpatients with AECOPD were prospectively enrolled from seven medical centers in China between December 2018 and June 2020. On admission, D-dimer was detected, PPS was calculated for each patient, and the incidence of 2-month VTE was investigated. The receiver operating characteristic (ROC) curve was used to evaluate the predictive value of D-dimer and PPS on VTE development, and the best cut-off value for both methods was evaluated through the Youden index. Results: Among the 4468 eligible patients with AECOPD, 90 patients (2.01%) developed VTE within 2 months after admission. The area under the receiver operating characteristic curves (AUCs) of D-dimer for predicting VTE were significantly higher than those of the PPS both in the overall cohort (0.724, 95% CI 0.672-0.776 vs 0.620, 95% CI 0.562-0.679; P<0.05) and the subgroup of patients without thromboprophylaxis (0.747, 95% CI 0.695-0.799 vs 0.640, 95% CI 0.582-0.698; P<0.05). By calculating the Youden Index, the best cut-off value of D-dimer was determined to be 0.96 mg/L with an AUC of 0.689, which was also significantly better than that of the PPS with the best cut-off value of 2 (AUC 0.581, P=0.007). After the combination of D-dimer with PPS, the AUC (0.621) failed to surpass D-dimer alone (P=0.104). Conclusion: D-dimer has a superior predictive value for VTE over PPS in inpatients with AECOPD, which might be a better choice to guide thromboprophylaxis in inpatients with AECOPD due to its effectiveness and convenience. Clinical Trial Registration: Chinese Clinical Trail Registry NO. ChiCTR2100044625; URL: http://www.chictr.org.cn/showproj. aspx?proj=121626.