The United Nations 2030 Agenda catalysed the development of global target-seeking sustainability-oriented scenarios representing alternative pathways to reach the Sustainable Development Goals (SDGs). Implementing the SDGs requires connected actions across local, national, regional, and global levels; thus, target-seeking scenarios need to reflect alternative options and tensions across those scales. We argue that the design of global sustainability-oriented target-seeking scenarios requires a consistent process for capturing multiple and contrasting perspectives on how to reach the goals, including the perspectives from multiple scales (e.g. local, national, regional) and geographic regions (e.g. the Global South). Here we propose a novel approach to co-design global target-seeking scenarios, consisting of (a) capturing global perspectives on pathways to the SDGs through a review of existing global scenarios; (b) a multi-stakeholder process to obtain multiple sub-global perspectives on pathways to sustainability; (c) an analysis of convergences, and crucially, divergences between global and regional perspectives on pathways to reach the SDGs, feeding into the design of new target-seeking scenario narratives. As a case study, we use the results of the 2018 African Dialogue on The World in 2050, discussing the future of agriculture and food systems. The identified divergent themes emerging from our analysis included urbanization, population growth, agricultural practices, and the roles of different actors in the future of agriculture. The results challenge some of the existing underlying assumptions of the current sustainability-oriented global scenarios (e.g. population growth, urbanisation, agricultural practices), indicating the relevance and timeliness of the proposed approach. We suggest that similar approaches can be replicated in other contexts to better inform the process of sustainability-oriented scenario co-design across scales, regions and cultures. In addition, we highlight the implications of the approach for scenario quantification and the evolution of modeling tools.