This research was based on the data obtained from a questionnaire survey on the consumers' behavior around the Nanjing Road. In this paper, the survey area is divided into several blocks as study units. The consumers' basic behavioral characteristics including gender, age, occupation, living place, distribution of flow, stop times, money consumed as well as spatial phase characteristics (entrance, multistop, and exit) are analyzed. Then the multinomial logit model is calibrated to explain the spatial choice behavior of consumers. The model gets satisfactory results and reveals how physical environment in the commercial space influences consumer behavior. The multistop track is derived by model simulation, which helps to find some behavioral characteristics and problems in spatial usage. Finally, behavior changes are predicted and evaluated based on the model.