The solder joint is difficult to position accurately during the resistance spot welding process of enameled wire,resulting in welding quality problems such as virtual welding,weld misalignment and lack of weld. In order to solve these problems, assembly recognition and positioning system for resistance spot welding of enameled wire based on machine vision is developed. The various kinds of bad patterns during the resisance spot welding positioning process of enameled wire are analyzed,the image processing algorithms of recognition and positioning system are studied,and the strategy and rules of positioning are formulated. Firstly,the region images of enameled wire and pad are acquired by extracting ROI with template matching;secondly,the bad patterns are recognized and eliminated by means of image processing algorithms such as thresholding,morphology and Laplacian operator;lastly,the optimum solder joint position are computed through extracting the edge of enameled wire and pad. The experiment results show that the positioning system can effectively eliminate the bad patterns, accurately extract and output the coordinates of the optimum solder joint position. This system can meet the requirements of automatic resistance spot welding of enameled wire because its recognition and extraction time is very short.