Due to properties of flexible material, uneven thickness, cutting speed, rapid changing feed direction and other factors in path processing of flexible material, the edge of processing path image become fuzzy, causing difficulty of improving edge detection precision. In this paper, after analyzing characteristics path contour of flexible material, taking gray-level integration of target area as a regional energy, and adding it to traditional Snake model which is widely used in edge extracting of complex geometric shapes, combining image edge with region information to build a new Region Snake (R-S) model, therefore by certain transformation, the description of total energy of contour can be replaced by force acting that moving on the contour. In new R-S model, matrix expression of discrete force balance equation of contour is obtained by discretization of finite difference method, then derivative terms are introduced and to get the calculation formula of processing path contour curve through Euler iteration method; finally by judging local extremum of covariance matrix of curve to realize corner detection. Three patterns with different shape characteristics were used to test the R-S model, Experimental results show that, the target contour curves of three patterns are completely extracted, the mean square errors of extracted curves are less than 0.35%, the detection rates of corners are greater than 90%, these prove that the R-S model proposed for flexible materials processing path contour extraction has high calculation accuracy and good stability, the method of corner detection is very suitable for flexible material processing path contour.