The recognition of truck cab assembly identifier is an important working procedure in the completed car assembling. Many different truck models will be assembled in the completed car assembling. It is necessary to recognize the type of truck cab assembly before assembly. An auto-inspection system for truck cab assembly identifier based on computer vision technology has been developed. The identifier image will be gotten by CCD camera by using of LED lights. In order to suit production rhythm of assembly line, SSDA (sequential similarity detection algorithm), quick edge threshold arithmetic and direction statistic eigenvector algorithm have been adopted to obtain the characters of picked image. Thinking of less identifier types and less varieties, direct segmentation of character will be realized by using vertical and horizontal projection. A standard template can be built, considering that the character congregation is small. Minimum distance Recognition has been applied to identify characters. It has proved that the system based on VC++ and Matlab can recognize all kinds of truck cab assembly identifier efficiently in the condition of ensuring production rhythm.