A two-year field experiment was conducted to investigate the potential application of soil electrical conductivity for variable rate seeding. Map calculations and geo-statistical techniques were used to establish a relationship among EC, elevation, seeding rate, and yield. Yield data taken from the first year, which was a dry year, showed inconsistent relationships and the overall yield variation within the farm was very low at 700 kg ha. However, yield data from the second year, being a wet year, exhibited relationships where an increase in EC indicated an increased yield potential, while increased seeding rates exhibited fluctuating trends in yield potentials. There is clear evidence that the existing relationship between the site properties as quantified by the EC, the seeding rate, and the crop yield, can favorably be used for Variable Rate Planting (VRP) in this particular production system. Regarding the development of prediction models for use in these situations, linear and non-linear parametric models were tried on the 2003 data, but with little success. A generalized additive model, a non-parametric approach, was used next and the yield model developed was found to regress the relationship adequately. To further improve the prediction accuracy, a Neural Network (NN) technique was used on the data. The diagnostics indicated that the yield estimation was precise (R = 0.89). The NN approach was identified as a very promising technique for using EC data in the successful application of variable rate technology towards maximizing yield potential of sites. However, the results indicate that the modeled relationship is specific to only that particular crop production system.