The aim of the present study was to determine the diagnostic accuracy of computed tomography (CT)-guided core needle biopsy (CNB) and to retrospectively analyze the correlation between the factors and complications of the procedure. Between January 2009 and June 2010, CNB was performed on 345 lung lesions in 343 patients. These patients were then followed up for at least two years. The sensitivity, specificity, accuracy, positive predictive value (PPV) and negative predictive value (NPV) of the CNB diagnoses were calculated. The correlation between factors, such as smoking, positoin and maximal diameter, and the complications of pneumothorax and hemorrhage was analyzed by χ2 test. The sensitivity, specificity, accuracy, PPV and NPV of the CNB diagnoses were 97.3, 100, 97.7, 100 and 87.7%, respectively. A statistically significant correlation was found between pneumothorax and the factors of smoking (P=0.015) and position (P<0.01) and length of the needle in the normal parenchyma (P=0.011), as well as between hemorrhage and the maximal diameter (P=0.005) and length of the needle in the normal parenchyma (P<0.01) and the frequency of needle adjustments (P<0.01). A CT-guided core needle biopsy of the lung lesions provides a high diagnostic yield. Smoking, the decubitus position and a longer length of the needle in the normal parenchyma were found to represent risk factors for a pneumothorax. In addition, a small diameter and longer length of the needle in the normal parenchyma and a more frequent adjustment of the needle were poor predictive factors of hemorrhage.