Abstract: Objective To evaluate the predictive value of improved smoking index in patients with arteriovenous thromboembolic disease after thoracoscopic radical resection of lung cancer and predict the optimal diagnostic value. Methods A total of 123 patients who were diagnosed with primary lung cancer and treated with thoracoscopic radical resection from January, 2015 to November, 2019 were enrolled and their clinical data were retrospectively analyzed. According to the occurrence of thromboembolic disease during the perioperative period, the patients were divided into two groups:a thromboembolic group (n=35, including 13 cases of venous embolization and 22 cases of arterial embolization) and a non-thromboembolic group (n=88). Both groups were compared for age, sex, body weight index (BMI), tumor type and classification, past history, alcohol consumption, chronic obstructive pulmonary disease (COPD), the length of anesthesia, time in bed, type of surgery, improved smoking index and interleukin (IL)-6 levels. A receiver operating characteristic (ROC) curve was plotted to evaluate the value of improved smoking index and IL-6 in predicting thromboembolic disease. Results Compared with the non-thromboembolic group, the thromboembolic group showed remarkable increases in the length of anesthesia, improved smoking index and IL-6 levels (P<0.05). There were no differences in other factors. Multivariate logistic regression analysis indicated that improved smoking index, length of anesthesia, hypertension and age were independent risk factors for postoperative thromboembolic disease after thoracoscopic surgery. Furthermore, the AUC of improved smoking index for predicting arterial embolization was 0.903 (95%CI: 0.831-0.951), with the optimal cut-off value of 26.44, and a predictive sensitivity of and 77.27%, and a specificity of 93.18%. In contrast, the AUC of improved smoking index for predicting venous embolization was 0.780 (95%CI: 0.687-0.856), with an optimal cut-off value of 26.44, a predictive sensitivity of 53.85% and a specificity of 93.18%, respectively. The AUC of IL-6 for predicting thromboembolic disease was 0.917 (95%CI: 0.853-0.959), with an optimal cut-off value of 0.062 8 μg/L, a predictive sensitivity of 94.29% and a specificity of 78.41%. Conclusions The combined use of improved smoking index and IL-6 can be adopted to predict arteriovenous thromboembolic disease after thoracoscopic radical resection of lung cancer, which will be supportive for early prevention and treatment of thromboembolic disease, with important significance for improving the prognosis of these patients.