Comparing the impact of different ways to divide Traffic Analysis Zones (TAZs) on traffic assignment can provide a general method, which helps to find a better solution with less costs, for urban planning process and transportation policy making. Previous literature has mostly studied foreign metropolitans but overlooks smaller-scaled regions. Thus, based on small and medium-sized cities, this paper uses quantitively methods to figure out the results of impact that TAZs division on traffic assignment. Zhoukou (located in Henan province) is selected as the object. Using Smith sampling model, the travel data table was obtained by GPS extraction with questionnaire survey. Seven different ways to divide TAZs is considered and evaluated using TransCAD respectively. It is found that the effectiveness and accuracy of traffic assignment increase with the number of TAZs increasing, but there is a “threshold”, which means the results would only fluctuate within a certain range if subdividing TAZs continually. However, when dividing TAZs with triangles and squares, traffic assignment results are similar, and they are superior to the results obtained by using traditional division ways with the same division number. Under the premise that the research area is not large and the number of divisions is limited, utilizing the grid method to divide TAZs can reduce the workload and acquire a preferable traffic prediction result. For the possible deviation, the reasons are discussed including the limitation of samples and software.

CICTP