In the context of aquaculture, carrying capacity is generally understood as the standing stock of a particular species at which production is maximised without negatively affecting growth rates. The estimation of carrying capacity for aquaculture is a complex issue. That complexity stems from the many interactions between and among cultivated and non-cultivated species, as well as between those species and their physical and chemical environments. Mathematical models may help to resolve these interactions, by analysing them in a dynamic manner. Previous carrying capacity models have considered the biogeochemical processes that influence growth of cultivated species in great detail. However, physical processes tend to have been addressed very simplistically. Further, most modelling has been for monocultures, despite the increasing importance of multi-species (=polyculture) systems.We present here a two-dimensional coupled physical–biogeochemical model implemented for Sungo Bay, Shandong Province, People’s Republic of China. Sungo Bay is used for extensive polyculture, where bivalve shellfish and kelp are the most important cultivated species. Data collected over 13 years (1983–2000) was available for modelling. Our main objectives were to implement the model, achieving reasonable calibration and validation with independent data sets, for use in estimating the environmental carrying capacity for polyculture of scallops and oysters.Findings indicate that the model successfully reproduces some of the main features of the simulated system. Although requiring some further work to improve predictive capability in parts, predictions clearly indicate that Sungo Bay is being exploited close to the environmental carrying capacity for suspension-feeding shellfish. Comparison of different culture scenarios also indicates that any significant increase in yield will depend largely on a more optimal spatial distribution of the different cultivated species.