Objective: To evaluate the clinical value of Lung Vessels software in diagnosis of pulmonary embolism (PE) in dual energy CT pulmonary angiography (DE-CTPA). Methods: Totally 49 patients of suspected PE underwent DE-CTPA. Lung Vessels software was used to process the images and the location and number of the abnormal vessel were recorded. The sensitivity, specificity, positive predictive value and negative predictive value of Lung Vessels were calculated with CTPA combined with clinic as gold standard based on patient, segment and sub-segment numbers. Results: CTPA revealed 29 patients of PE with 188 diseased vessels, including 138 segments and 50 subsegments. The sensitivity, specificity, positive predictive value and negative predictive value of Lung Vessels in diagnosing PE was 96.55% (28/29), 45.00% (9/20), 71.79% (28/39), 90.00% (9/10) for patient as a unit, 65.94% (91/138), 80.38% (598/744), 38.40% (91/237), 92.71% (598/645) for segment as a unit and 64.00% (32/50), 92.24% (1581/1714), 19.39% (32/165), 98.87% (1581/1599) for subsegment as a unit, respectively. Conclusion: In DE-CTPA, Lung Vessels software can directly show the lung vessels, which is helpful for diagnosing PE, especially segmental and subsegmental PE with potential application prospects. Copyright © 2014 by the Press of Chinese Journal of Medical Imaging and Technology.