There are five edge detection techniques: Sobel Method, Roberts Method, Prewitt Method, Canny Method and Laplacian of Gaussian Method. An image of each of the fourteen species of orchid were selected, and encoded to the four types of image extension: Joint Photographic Experts Group (.JPEG), Portable Network Graphics (.PNG), Graphics Interchange Format (.GIF) and Bitmap Image File (.BMP). Each of the fifty-six images was processed with each edge detection technique using MATHLAB R2008B. To compare the edge detection technique on each image, the root mean square (RMS) of each image is used. The results show that the RMS of the Canny method on the .JPG and .GIF images was less than the other methods (5.24E+05, 4.37E+05). In the case of the .PNG images, the Prewitt method was the most suitable method with a RMS of less than the other four methods. The Robert method was the most suitable for .BMP. Further study will be undertaken to implement a system for identification of these Thai Paphiopidelum orchids from images.