Remote sensing in the form of aerial color infrared (CIR) photography has been shown as a tool for in-season N management in winter wheat (Triticum aestivum L.). The objectives of this study were (i) to develop a methodology for predicting in-season N requirement for winter wheat at growth stage (GS)-30 using aerial CIR photography, and (ii) to determine if the relationship between optimum N30 rates and spectral indices measured at GS-30 is sensitive to the density of the wheat crop (biomass). Field studies were conducted for 3 winter wheat growing seasons over a wide range of soil conditions across North Carolina. Different planting date seeding rate (PDSR) combinations were applied to create a range of biomass levels at GS-30. Different levels of N were applied (i) at GS-25 (N25) to create a range of N supply and winter wheat color and near infrared (NIR) radiance and (ii) at GS-30 (N30) to measure grain yield response to N30. Aerial CIR photographs were obtained at each site at GS-30 prior to N applications. Significant biomass response to PDSR and yield response to N25 and N30 was observed. Optimum N30 ranged from 0 to 124 kg ha-1 with a mean of 55 kg ha-1. Better prediction of optimum N30 rates was obtained with spectral indices calculated relative to high-N reference strips compared to absolute bands or indices. Biomass measured at GS-30 influenced the strength of the relationship between optimum N30 and spectral indices. The best predictor of optimum N30 (R2 = 0.85) was a quadratic model based on measured winter wheat radiance in the green (G) band expressed relative to high-N reference strips (Rel GS) when the GS-30 biomass was > 1000 kg ha-1.