Lung segmentation operation is handling a very important role for determining unhealthy tissues and thus diagnosis and treatment of lung cancer. In this work Lung CT images are segmented by using Random Walk algorithm. First 25 DICOM images belong to 4 patients are read for processing and 8-bit gray level images obtained. For interactive segmentation, the seed points are set and applied as input parameters to the algorithm. Using the output values, the segmentation finished with an accuracy of average %95.