A long-term study was conducted on nitrogen application rate and its impact on soil residual nitrogen and cotton (FM960B2RF) lint yield under a drip irrigation production system near Plainview, Texas. The experiment was a randomized complete block design with five nitrogen application rates (0, 56, 112, 168 and 224 kg per ha) and five replications. The soil nitrogen treatment was applied as side dressing. Cotton yield, leaf N, seed N, soil residual nitrate, amount of irrigation, and rainfall data were measured per growing season for 8 years (2002-2009). The latitude, longitude and altitude of four corners of each plot were recorded with GPS. Nearest neighbor distance from each nitrogen treatment plot and the north-south slope and east-west slope of each plot were calculated. The data matrix was analyzed with principal component analysis to determine the contribution of each variable to total variation in the data. After seven years of continuous application of variable rate of N, residual N levels varied significantly between the 224 kg/ha treatment and lower nitrogen level (0 and 56 kg/ha) treatments. Variation in residual N levels coupled with variable N application resulted in phenotypic expression of nitrogen deficiency in cotton across treatment plots, especially between zero-N plots and N-applied plots. Lint yield varied with nitrogen level as expected. The soil nitrate residuals in the control plot on each succeeding year did not decline significantly and showed a significant rise in 2008. The study provided the evidence of high N treated plots influencing residual N in low N or untreated plots which could be useful in precision nitrogen fertilizer application in cotton.