In this paper, the algorithm of mixed Gauss background modeling is analyzed from two aspects of theory and application. Accurate and robust vehicle detection and the recognition still a challenging task in the field of intelligent transportation surveillance systems. This exhibits promising potential for implementations with real-world applications.