Aquaculture is the world’s fastest growing food production sector (FAO, 2007). The continuous growth in many countries, however, relies heavily on the ability to reduce the emission of nutrients and chemicals from the fish farms. A way to manage and treat the nutrient aquaculture wastes is by production in recirculating aquaculture systems (RAS). In Denmark, more than 50 % of total fresh-water rainbow trout production is made in semi-intensive RAS, called ModelTroutFarms (MTF). MTF efficiently removes organic matter (93%), phosphorous (76%), and nitrogen (50%) (Svendsen et al., 2008). This makes nitrogen the limiting process parameter for further environmentally viable increase in production. Nitrogen removal is a two step transformation process, with (1) ammonia-N oxidation to nitrate-N in the RAS’ biofilter, and subsequently (2) nitrate-N reduction to N2 in the constructed wetlands. The latter being the final cleaning component of the MTF set-up. No specific denitrification filter has so far been implemented in Danish MTFs. An in-situ study was conducted at a commercial MTF (1000 ton/year) for evaluating the potential of using the fermentation products from anaerobic digestion in the sludge storage basins, to fuel denitrification in specific denitrification filters. In experimental filters (5.5 m3) nitrate-containing outlet water was mixed with drainage water from the sludge storage basins according to a factorial design varying C/N ratio from 4 to 12 (CODs /NO3-N) and hydraulic retention time (HRT) from 50 to 180 min. The highest removal rate recorded, 125 g NO3-N/m3reactor/d, was found in treatments at the design center point, and multivariate response surface analysis modeled a maximum N-removal at C/N ratio of 8.8 and HRT of 114 min. The effect of C/N ratio depended on the HRT: At low HRT, variation in C/N ratio had no effect on N-removal. On the contrary, at high HRT, the highest N-removal was measured at high C/N ratio but significant ammonia-N was simultaneously produced, most probably by dissimilatory nitrate reduction to ammonia (DNRA). Running the filters at high HRT and low C/N ratio rendered a relatively lower nitrate-N removal rate but significantly higher ammonia-N reduction, which could indicate anaerobic ammonia oxidation (anammox) activity. A controlled laboratory anaerobic MTF sludge digestion experiment showed that app. 40% additional nitrate-N reduction could theoretically be achieved if implementing the use of fermented sludge as carbon source for denitrification. Besides the N-reduction, the directly linked sludge (/organic matter) reduction is a beneficial side effect of such an operational set-up