Coastal social–ecological systems (SES), particularly in large bays, are critical for fisheries, transportation, and disaster prevention in island and coastal countries. To achieve the sustainability of such bays, public involvement is recently considered inevitable for planning and management, but the increasing complexity of variables and future visions to be considered is one difficulty when trying to include many stakeholders and public opinions. To address this challenge, a free-associative description questionnaire survey was used in this study to extract holistic coastal residents’ future visions for Tokyo Bay, including both positive and negative outcomes. By integrating biterm topic modeling (BTM) and multiple-factor analysis (MFA), this study succeeded to aggregate and visualize the various future visions of Tokyo Bay with enhanced comprehensibility. As one outcome, the linkages and differences between the major topics in the positive and negative future visions were visualized as vectors in a correlation circle. Also, the study found that these two kinds of future vectors are not always polar opposites, but, rather, some of them are interlinked, pointing in the same direction. This highlights the importance of measuring the balance between two kinds of future vectors in consensus-building in order to search for the optimal future direction. Finally, the study discusses the potential of this method as a “Future Compass”, for implementing future-oriented consensus-building toward the sustainability of SES.