Potato (Solanum tuberosum, L.) yield and quality are highly dependent on the availability of nitrogen (N) during the crop’s critical growth stages. Canopy-level hyperspectral (HS) imagery has been shown to be an effective research tool for determining the best wavelengths and/or indices for the detection of N stress in a number of crops. Research findings from HS imagery can be applied to active sensors as a way to increase the effectiveness of real-time, variable rate N applications, but limited data exist for potato production. A field study was conducted in 2010 and 2011 at the Sand Plain Research Farm in Becker, MN on a Hubbard loamy sand soil to evaluate the effects of water stress, N fertilizer rate/timing, the varieties Russet Burbank (RB) and Alpine Russet (AR), and growth stage on the ability of canopy-level HS imagery (401-982 nm) to detect N stress in a potato crop. The ability of canopy-level narrowband reflectance to detect differences in potato N status was evaluated by performing linear regression between HS narrowband reflectance and leaf N concentration as a means to distinguish the optimum wavelengths to detect N stress. As the potato crop matured, the coefficient of determination (r²) decreased at most HS wavelengths, especially in the NIR region. The far green and near red regions from ~582-610 nm, and those at the beginning of the red-edge from ~685-695 nm were the best performing wavelengths for detecting N stress overall. Canopy reflectance was able to predict leaf N among image dates more consistently for RB than for AR.