Given the expansion of urban agriculture (UA), we need to understand how this system provides ecosystem services, including foundational societal needs such as social cohesion, i.e., people's willingness to cooperate with one another. Although social cohesion in UA has been documented, there is no framework for its emergence and how it can be modeled within a sustainability framework. In this study, we address this literature gap by showing how the popular cascade ecosystem services model can be modified to include social structures. We then transform the cascade model into a bottom-up causal framework for UA. In this bottom-up framework, basic biophysical (e.g., land availability) and social (e.g., leadership) ecosystem structures and processes lead to human activities (e.g., learning) that can foster specific human attitudes and feelings (e.g., trust). These attitudes and feelings, when aggregated (e.g., social network), generate an ecosystem value of social cohesion. These cause-effect relationships can support the development of causality pathways in social life cycle assessment (S-LCA) and further our understanding of the mechanisms behind social impacts and benefits. The framework also supports UA studies by showing the sustainability of UA as an emergent food supplier in cities.