Abstract: The urban cognitive map analysis methodology which Kevin Lynch pioneered has been widely used, but the research on spatial image in small towns still needs to be expanded. The cognitive map research methodology is employed in this study and seven small towns of Yantai City region are selected as the cases. A total of 105 valid cognitive maps were collected, and then the constituent elements and characteristics of the image in small towns would be analyzed. The preliminary research results are as follows:(1)The cognitive components of small towns' image are primarily composed by paths and landmarks, while the other elements, such as districts, edges and nodes, are relatively not significant. The occurrence frequency of paths and landmarks is connected with road network in the township and the function of landmarks. (2)Residents trend towards confusing the township interface, namely ‘village-township-town mix’, which is influenced by the size of townships and so on. (3)Among the cognitive types, the proportion of sequential maps and low-level cognitive types is higher. (4)The comprehensive cognitive orientation of residents in small towns is mainly in the north-south direction, which is related to road network and main streets' orientation. The analysis of collective cognition shows that large-scale development projects like real estate development have impacts on the perception of residents in small towns. Three types of collective images of small towns are summarized:town-village integration, town-village independent, and enclave development type. It is hoped that the interpretation of spatial image in small towns in this article can provide a reference for the construction, planning and urban-rural integration of small towns in China.