In this paper, we propose an automatic lung nodule segmentation using intensity, curvature and morphology information in CT images. Our method consists of following three steps. First, a nodule type (isolated, juxtapleural, juxtavascular or pleural tail nodules) is determined by calculating a boundary coverage ratio of the nodule region in each slice. Second, an initial nodule segmentation using 3D seeded region growing is performed to extract a nodule candidate which includes the nodule and its neighbor structures. Third, the nodule candidate is refined by applying different algorithm according to nodule type for separation neighbor structures such as chest wall and vessels. For the juxtapleural nodule, connections between the nodule and chest wall are separated by linking two feature points with great curvature of the nodule candidate boundary. For the juxtavascular nodule, connections between the nodule and vessels are separated by performing intersection operation using circular structuring element with compact size in the nodule. For the pleural tail nodule, the segmentation algorithms of juxtapleural and juxtavascular nodules are successively performed. To evaluate the result of our method, we perform the visual inspection and accuracy evaluation. The experimental results show that nodules in CT images can be accurately segmented without leakage to the neighbor structures by applying an adaptive algorithm after determination of nodule type.