Abstract: In order to describe the shape features of tobacco strips accurately, by combining machine vision device with digital image analysis, an analysis method was developed, wherein the shape of tobacco strip was characterized by rectangular degree, compactness and elongatedness. The shape features of large, middle and small size strips and scraps in random strip samples were compared. The results showed that:1) Except for the influence of the mechanical deformation for tobacco strips in the process of image acquisition, the relative error and relative standard deviation of the developed method were less than 5.0%. 2) The Euclidean distance between large and middle size strips and standard rectangular strips was the shortest, it suggested that the shape of large and middle size strips was close to rectangle. 3) The shape of small size strips was close to standard square strips, that of scraps was close to standard circular strips. The relative standard deviations and relative errors of Euclidean distance of individual strip sample relative to five sets of standardized strips were all less than 5.0%, which indicated that the developed method offered consistent and reliable results.