A series of bench-scale experiments was performed with a novel electrodeionization (EDI) process to determine its capacity to remove NO3- from water in the presence of competing species. Experiments were performed with deionized water that was dosed with selected anionic and cationic species and with two natural waters that contained differing concentrations of dissolved solids. Extended duration experiments were performed to assess the ability of the process to sustain NO3- removal over a longer period of time. It was found that the use of ion-selective resins in the EDI unit resulted in a process that was more nitrate selective and energy efficient that when a general anion-exchange resin was employed. Operation in EDI mode was found to be superior to operation without resins (ED mode). The most highly selective and energy-efficient operation was at the lowest applied voltage of 2 V per membrane pair. The impact of competing species on NO3- flux was independent of the competing species when expressed on an equivalence basis. At elevated concentrations of multiple competing species the flux of NO3- was reduced. The removal of NO3- from two natural water streams revealed the significance of the water composition on the performance of the EDI unit with respect to the selectivity of NO3- removal and energy consumption. When treating the Manotick well water, the EDI unit required more energy per equivalence of NO3- removed and was less selective than when treating the Ottawa potable water that was lower in dissolved solids. Extended duration experiments with the dosed Ottawa water over a period of 48 h did not reveal any decline in fluxes.