Against the background of e-commerce, online shopping has seen considerable growth in China, as in the rest of the world. The last mile of delivery services for online shopping is a logistics challenge that affects service performance. This study has two main aims: first, to construct an evaluation criteria system for last-mile delivery service; and second to propose a matching model, capable of ranking six delivery methods according to customer preferences in the different urban areas. The factors base is established from the literature and from questionnaires and interviews with experts. Moreover, by conducting a questionnaire with consumers and analyzing the data, this research identifies the top 15 factors. The matching model based on the fuzzy analytic hierarchy process is constructed to compute the weight of each factor. The collection of data on customer preference was performed in distinct urban areas. Finally, to illustrate the validity of the criteria factors and the matching model, it was applied to three districts in Chongqing in China. Finally, our theoretical results from the experiments in real-life instances show that the criteria system and the matching model could help express companies to identify appropriate delivery methods in specific areas.