To explore the diagnostic efficiency of percutaneous lung biopsy for malignant lung tumors under the guidance of computed tomography (CT). The clinical data of 623 patients undergoing CT-guided percutaneous lung biopsy at Department of Respiratory Medicine, Xinqiao Hospital, Third Military Medical University from January 2010 to November 2012 were analyzed to examine the influencing factors for the occurrences of pneumothorax and hemoptysis, including patient age, gender, site of puncture, tumor location and size, depth and number of puncture, chronic obstructive pulmonary disease (COPD) and doctor's clinical experience. A total of 623 cases were recruited. There were 410 male and 213 female with an average age of 60 ± 11 (15-87) years. Among them, 618 were diagnosed as malignant lung tumors with a positive rate of 99.2%. Pathological classifications were adenocarcinoma (n = 387, 62.1%), squamous carcinoma (n = 166, 26.6%), small cell lung cancer (n = 30, 4.8%), undifferentiated cancer (n = 18, 2.9%) adenosquamous carcinoma (n = 2, 0.3%), sarcomatoid carcinoma (n = 6, 0.9%), leiomyosarcoma (n = 3, 0.5%), neuroendocrine carcinoma (n = 2, 0.3%), low differentiated carcinoid (n = 1, 0.2%), cribriform carcinoma (n = 1, 0.2%), malignant lymphoma (n = 1, 0.2%) and metastatic renal cell carcinoma (n = 1, 0.2%) respectively. Another 5 patients (0.8%) had an initial diagnosis of chronic inflammation according to their pathological features of lung punctures. Their definite diagnoses were lung squamous cell carcinoma (n = 3), adenocarcinoma (n = 1) and small cell carcinoma (n = 1) respectively. The major complications were pneumothorax and hemoptysis. Among 73 cases of pneumothorax (n = 73, 11.7%), 65 cases (n = 65, 10.4%) of mild pneumothorax (lung compression < 20%) were cured by sufficient oxygen inhalation and repose while 8 cases (1.3%) suffered lung compression ≥ 75% recovered via closed thoracic drainage. Eighty-four patients (13.5%) suffered from the complications of slight or self-limited hemoptysis with blood-tinged sputum. The symptoms of hemoptysis disappeared after oral hemostatic or other treatments. Another 5 patients (0.8%) suffered from moderate hemoptysis (100-500 ml) and their symptoms disappeared after rest or an intravenous infusion of haemostatic. None of them died of massive hemoptysis or biopsy. The occurrence of pneumothorax was significantly correlated with the diameter of lesions ( ≤ 2 cm) (P = 0.006), depth of needle (>3 cm) (P = 0.044), number of puncture needles ( ≥ 3 needles) (P = 0.025), inadequate clinical experience of puncturing doctors (P = 0.003) and COPD (P = 0.036); the occurrence of hemoptysis was obviously correlated with the diameter of lesions ( ≤ 2 cm) (P = 0.031), depth of needle (>3 cm) (P = 0.018), number of puncture needles ( ≥ 3 needles) (P = 0.015) and inadequate clinical experience of puncturing doctors (P = 0.014). However, the relationship was unremarkable between the occurrences of pneumothorax and hemoptysis and patient gender, age, sites of puncture and tumor location. With a high diagnostic value, CT-guided percutaneous lung biopsy is both safe and effective for malignant lung tumors.