Understanding factors controlling primary production is fundamental for the protection, management, and restoration of ecosystems. Tropical seagrass ecosystems are among the most productive ecosystems worldwide, yielding tremendous services for society. Yet they are also among the most impaired from anthropogenic stressors, prompting calls for ecosystem‐based restoration approaches. Artificial reefs (ARs) are commonly applied in coastal marine ecosystems to rebuild failing fisheries and have recently gained attention for their potential to promote carbon sequestration. Nutrient hotspots formed via excretion from aggregating fishes have been empirically shown to enhance local primary production around ARs in seagrass systems. Yet, if and how increased local production affects primary production at ecosystem scale remains unclear, and empirical tests are challenging. We used a spatially explicit individual‐based simulation model that combined a data‐rich single‐nutrient primary production model for seagrass and bioenergetics models for fish to test how aggregating fish on ARs affect seagrass primary production at patch and ecosystem scales. Specifically, we tested how the aggregation of fish alters (i) ecosystem seagrass primary production at varying fish densities and levels of ambient nutrient availability and (ii) the spatial distribution of seagrass primary production. Comparing model ecosystems with equivalent nutrient levels, we found that when fish aggregate around ARs, ecosystem‐scale primary production is enhanced synergistically. This synergistic increase in production was caused by nonlinear dynamics associated with nutrient uptake and biomass allocation that enhances aboveground primary production more than belowground production. Seagrass production increased near the AR and decreased in areas away from the AR, despite marginal reductions in seagrass biomass at the ecosystem level. Our simulation's findings that ARs can increase ecosystem production provide novel support for ARs in seagrass ecosystems as an effective means to promote (i) fishery restoration (increased primary production can increase energy input to the food web) and (ii) carbon sequestration, via higher rates of primary production. Although our model represents a simplified, closed seagrass system without complex trophic interactions, it nonetheless provides an important first step in quantifying ecosystem‐level implications of ARs as a tool for ecological restoration.