In recent years, researchers have been studying spectral reflectance for estimating N nutrition of crops and the effect of spatial scales in making measurements related to N research. The application of indices developed at the leaf scale is often unsatisfactory for assessing N content at the canopy scale. In 2010 and 2011, experimental studies were conducted with the objectives to compare reflectance measurements of cotton grown at different N levels, and to estimate N content of field-grown cotton using a model based on remote sensing. Reflectance measurements were made using a spectroradiometer at different spatial scales, i.e., leaf, canopy (excluding soil), and scene (analogous to satellite or airborne observations). Since N-deficiency primarily affects cotton physiological characteristics such as the leaf greenness and biomass, it was expected that the low N rate would consistently produce higher green reflectance and lower NIR reflectance, regardless of spatial scale, in comparison to the high N rate. Preliminary results, however, showed that the reflectance measurements made at the leaf and canopy scales did not support this expectation. Nevertheless, the scene measurements successfully depicted the expected results. The lack of consistency in reflectance measurements may explain previously unsuccessful attempts to develop indices at one scale and apply them at another scale. The consistency of the scene reflectance for different N rates was affected by the percent ground cover. Therefore, ground cover has been included as a factor in the model. The development of this model emphasizes the partial canopy chlorophyll content rather than leaf concentration. This concept was adopted since leaf nutrient data collection is commonly confined to a point basis, and better algorithms are required for estimating N at the landscape scale. Also, preliminary results indicated that the ground cover factor was shown to be a better estimator of crop biomass than NIR reflectance. Many researchers have found that high N treatments often have a NIR plateau in reflectance that is lower than that for intermediate N rates. Thus, it is expected that the model proposed in this study will have general applicability for estimating canopy chlorophyll content of cotton and replace biomass estimated from NIR reflectance with the more relevant estimation based on ground cover.