The awareness about environmental complexity involves real-time knowledge and demands urban planning initiatives. Knowledge is multiform, multi-agent and mirrors environmental complexity. Problems characterizing urban sustainability particularly claim non-expert knowledge, being informal, puzzling, uncertain, incomplete, hard to be handled, formalized, modelled. This study utilizes Fuzzy cognitive maps to explore such complexity and support multiagent decisions. It concerns the scenario-building process of the new plan of Taranto (Italy), a paradigmatic example of decaying industrial area, heavily characterized by social fragmentation and environment degradation. This approach aims at structuring environmental problems, modelling future strategies and contributing to build a multi-agent decision support system for complex urban planning contexts.