The width of cut tobacco has significant influences on the physical properties and sensory qualities of cigarettes. Due to the narrowness, complex distribution and different widths properties of the cut tobacco, a width measuring method is proposed based on adaptive image processing technologies. The high definition image of cut tobacco is attained by the scanning of highly numerous pixel linear charge coupled device (CCD). Widths are measured by applying series of image processing courses, such as binarization, image segmentation, decontamination, refining, edge extraction and matrix turning. Applying this method on actual cut tobacco images, a well adaptability to the complex distribution of the cut tobacco is verified. Additionally, accuracy is confirmed by contrasting the measurement results with standard template that fabricated from E-Beam lithographic tools.