Wetland loss along the Louisiana Gulf coast and excessive nitrate loading into the Gulf of Mexico are interrelated environmental problems. Nitrate removal by soil denitrification activity was studied in a ponded freshwater marsh receiving diverted Mississippi River water for the purpose of reversing or slowing wetland loss. Labeled ¹⁵N-nitrate was applied at 3.85 g N m^-2 into replicate study plots after removing above ground vegetations. Nitrogen gas (N₂) and nitrous oxide (N₂O) emissions from the plots were determined by isotope ratio mass spectrometry (IRMS). Emissions of ¹⁵N₂O were compared with the results determined by gas chromatography (GC). Results showed that it required 2 weeks to process the added nitrate with N₂O emission occurring over a period of 4 days. The apparent denitrification dynamics was assumed to follow the Michaelis-Menten equation. The maximum denitrification rate and Km value were determined as 12.57 mg N m^-2 h^-1, and 6.46 mg N L^-1, respectively. Therefore the maximum capacity of nitrate removal by the marsh soil would be equivalent to 110 g N m^-2 yr^-1, with more than 30% of nitrogen gas evolved as N₂O. For typical nitrate concentration in the Mississippi River water 1 mg N L^-1, nitrate would be removed at rate of 14.72 g N m^-2 yr^-1 with N₂O emission approximately 1.5%. A denitrification dynamic model showed that the efficiency of nitrate removal would largely depend on the water discharge rate into the ponded wetland. Higher discharge rate will result in less retention time for the water in the marsh where nitrate is denitrified.