Human eyes acquire numerous information and analyze an environment from day to day. Vision sensor is similar to that of human eye. One of research areas of vision sensor is driving control. The role of vision sensor is essential to control driving by controlling system such as automatic car and autonomous guided vehicle not human. Color recognition of the system such as automatic car is important so as to analyze a road sign like as traffic lights. RGB, which consists of red, green and blue, comes generally up when making a division of color in real world. However, it is difficult to recognize color by only RGB information. To recognize color landmarks for driving control, this paper takes RGB and CIE LAB color space into account. By composing these color system, color is determined by fuzzy inference system which is familiar to the way of human thinking. Circular Hough transformation helps to reduce data of an image effectively by extracting region of interest. In sequence, find the highest values about each of distribution maps. The highest values are converted to CIE LAB color space. To analyze and determine color of ROI, fuzzy inference system carry out by inputs which are original R, G, B components and a, b components.