The initial purpose of our study was to understand the preferences of stakeholders in green infrastructure for flood control using a discrete choice experiment. The discrete choice experiment makes it possible to assess individual preferences by asking respondents to choose preferred one among several scenarios consist of multi-attribute. According to economic utility theory, an inexpensive green infrastructure scenario should have been chosen under ceteris paribus conditions, but our results did not. From economic theory, inconsistent like ours are often interpreted as indicating bias and/or questionnaire design issues. However, we tried to give the results a different interpretation from perspective of relational values. We studied a case of green infrastructure in a large-scale flood control basin in Northern Japan. We conducted a questionnaire survey including choice tasks of the discrete choice experiment to town residents in Naganuma as stakeholders. The focus of our survey was to elicit the preferred green infrastructure scenario from a group of alternatives involving three ecosystem services (ecological status of the red-crowned crane, wetland area, establishment of bird-watching station). The respondents seemed to understand the scenario properly and valued the ecosystem services in a rational manner. But, as mentioned above, several aspects of the results of the discrete choice experiment remain to be interpreted appropriately. Therefore, we interpreted them using relational values. With the results of choice and membership parameters, we interpreted the results assuming that individual identity and place attachment, which are types of the relational values, are taken into consideration in the choice situation of a discrete choice experiment. We also found that social responsibility, which is also a type of the relational values, can help us to understand the unexpected results that cannot be interpreted in terms of economic utility theory. In choice situations which are related to the future vision of a region, relational value may have significant effect on them. Therefore, environmental valuation in such a situation requires a questionnaire design that assumes relational values and careful interpretation of its result with relational values.