Objective To test the validity of identifying the risk of venous thromboembolism (VTE) in severe inpatients with Padua risk assessment model. Methods In the retrospective case-control study, 78 inpatients with VTE (VTE group) and 96 inpatients without VTE (non-VTE group) admitted to the Intensive Care Unit (ICU) of Affiliated Hospital in North China University of Science and Technology from January 2015 to January 2017 were conducted in the same model. The risk of VTE was assessed according to the Padua risk assessment model, and the relationships between the scores and the dangerous levels were studied and compared between the two groups. The associations between the risk factors and VTE were analyzed with multivariate logistic regression analysis model. Results The average Padua risk scores in the patients with VTE were significant higher than those in the patients without VTE (P < 0.05). Compared with the patients with low scores, those with high scores were associated with 7.66-fold risk of VTE. From the logistic regression analysis on the relative factors in the severe inpatients, the most dangerous risk factor was bed rest of 3 or more days. Conclusions Padua risk assessment model can effectively predict the risk of VTE in severe inpatients.