Coastal and marine ecosystems are increasingly impacted by anthropogenic activities that act cumulatively with environmental changes, eroding biodiversity and its essential role in many ecosystem functions, including biomass production, which is vital for livelihood and food security. Understanding the links between socio-environmental factors, biodiversity, and fish biomass is essential for sustainable development. In this study, we use Structural Equation Models (SEM) to test, estimate, and explore complex relationships between these co-variates, distinguishing direct and indirect responses on five key coastal ecosystems across three ecoregions. In the Northwestern Mediterranean, fish biomass is positively linked to rocky habitats and taxonomic diversity but negatively influenced by sandy habitats. In the deeper rhodolith bed communities, taxonomic and functional richness, depth, and habitat coverage are the main correlates of fish biomass. In Northern Europe, sea surface temperature is a crucial factor directly influencing fish biomass, along with taxonomic richness, with depth and euphotic zone productivity playing significant roles. In the Lesser Antilles, taxonomic richness, depth, and live coral cover positively impact biomass, whereas terrace width has a negative effect, while no-take marine reserves boost taxonomic diversity. Across all ecosystems, taxonomic richness consistently enhances fish biomass, while entropy indices have negative impacts. Our results suggest that higher species diversity correlates with greater biomass, but that total biomass is dominated by a few species. Environmental factors, particularly habitat characteristics, outweigh direct anthropogenic influences like marine reserves. SEM reveal that, while pathways vary across ecosystems, taxonomic richness is a universal driver of fish biomass. Effective conservation and management strategies must consider these dynamics to mitigate human impacts and enhance the resilience of coastal ecosystems but also people livelihoods.