Objective To study the application value of a risk model based on drug metabolism-related genes in predicting the prognosis of lung cancer patients.Methods Based on the 298 pharmacokinetics (ADME)-related genes identified by Pharma ADME Consortium, ADME-related differentially-expressed genes (ADME-DEGs) were obtained through GEO chips GSE7670 and GSE32863. ADME-DEGs were analyzed with gene ontology (GO) functional enrichment and protein-protein interaction (PPI) network. By using consistent clustering and PCA principal component analysis, it was found that 18 ADME-DEGs could significantly divide lung cancer patients into two categories: Cluster1 (n=315) and Cluster2 (n=178). Through the LASSO algorithm, a risk model consisting of 10 ADME-DEGs was finally obtained. Through Kaplan-Meier survival analysis, ROC analysis and multivariate regression analysis, a Nomogram for predicting lung cancer patients was established based on the risk scores, and the predictive ability of the risk scores for the survival time of lung cancer patients was analyzed.Results A lung cancer risk prediction model was successfully established, constructed by 10 ADME-DEGs (SLC22A18, AOX1, CAT, ADH1B, SULF1, PPARG, CYP4B1, FMO2, GPX3, ABCA4). Kaplan-Meier survival analysis showed that patients with higher risk scores had significantly poorer overall survival. ROC analysis showed that risk scores could better predict the survival of lung cancer patients. Age, gender, and risk scores were significantly correlated with the overall survival of lung cancer patients. The Nomogram plot showed that the risk scores had a good predictive ability for the 10-year overall survival of lung cancer patients (C-index: 0.688).Conclusion The risk model constructed on the base of 10 ADME-DEGs had a potential role in predicting the prognosis of lung cancer patients after medication, and provided a new method for improving the prognosis of patients.