Abstract: A new automatic segmentation method of suspected lung nodules of edge-preserving smoothing and improved FCM algorithm is proposed.Firstly，the pulmonary parenchyma with smooth edge is separated from the original CT scans and the suspected lesions in the lung area are used by nonlinear anisotropic diffusion filtering for enhancement.Secondly，an improved Fuzzy C-Means（FCM） clustering algorithm is applied to the segmentation of the suspected lung nodules in different sizes.Experiment results indicate that the presented segmentation method of suspected lung nodules can be completed correctly and automatically，which achieves better performance than the conventional FCM algorithm and is much robust and efficient.