Lung cancer is one of the most common diseases and the leading cause of cancer death worldwide. In recent decades, the development of therapeutic drugs for lung cancer and the development of new biomarkers for the early diagnosis of lung cancer have made great progress. Although traditional preclinical models including in vitro cell line model have made great contributions to the development of targeted drugs and diagnosis kits, considerable shortcomings have attracted attentions due to uncoupled difference between preclinical and clinical results. Based on the principle of personalized medicine, patient-derived tumor xenograft (PDX) models are emerging as a novel approach for drug screening and new diagnosis agent finding. In this mini review, we summarized the establishment and characteristics of the PDX models and discuss their application in finding targeted drugs and diagnosis for treating lung cancer.