A simulation model was developed for an experimental recirculating mariculture system in flat, Israel. The system cultures gilthead seabream (Sparus aurata), using a seaweed biofilter (Ulva lactuca) to maintain water quality, reduce the flow-through rate of seawater, and reduce the nutrient load in effluents. The model describes the production, transfer, transformations, and losses of inorganic N in the system, in a relatively simple compartment model framework. The model was able to match very closely the experimental data on which it is based. Moreover, the model was able to predict the system's behavior in two additional experiments, which provided independent tests of the model under nonstandard operating conditions (reduced food input and reduced water recirculation rate). The success of the model beyond the range of conditions used in its construction suggests that the model provides a valid mechanistic description of the essential processes affecting inorganic N in the system. Sensitivity analysis of the model indicates that levels of ammonia-N and oxidized N in the system were most strongly affected by the rates of ammonia-N excretion by the fish and ammonia-N uptake by the seaweed, followed by the water inflow and recirculation rates, Sensitivities to internal transformation rate parameters were lower, which might explain the relative seasonal constancy of the system. Simulation experiments indicate that the experimental system could support much higher stocking densities of fish without exceeding tolerable levels of ammonia-N.