The paper makes some experiments on infrared image segmentation of ship object based on the fractal theory and MeanShift filtering.Using MFFK to pick out the region of interest(ROI) can remove large area background noise out of the ROI.Then,getting the binary mask of ROI will suppress the noise of the ROI further.Most of the noises maybe excluded via these two procedures.So using MeanShift filtering can get the distribution points of the target.At last,morphological filtering is adopted to obtain the object ship image.The experiment results show that the proposed algorithm can extract the ship object well.