This paper is intended to study the impacts of information guidance system on driving behavior at intersections. Using the structural equation model (SEM) and the improved technology acceptance model (TAM), a method was proposed to analyze the impacts on the driving behavior. On the basis of literature review, the factors that affect drivers' preference to the guidance information at intersections were determined. The factors can be divided into individual factors, system (or information) factors, and environmental factors. 1 544 valid samples were obtained with the preference survey of the drivers in Beijing, and the proposed model was examined and refined. Then, statistical methods were used to verify the hypothesis of the proposed model. The results have shown that perceived usefulness of the information, credibility of the information, and traffic conditions mainly affect the drivers' preference to the guidance information. The study provides supports to improve the efficiency of the information guidance system