For site-specific soil management in cotton (Gossypium hirsutum) production to be successful, soil property variation and their relationships with yield must be understood. The objectives of this study were to determine if soil factors commonly available to producers (soil fertility and texture data, elevation, and slope) could be related to cotton yield to make site-specific soil management decisions. Grid-based soil samples were collected from two fields and were analyzed for Lancaster-extractable Ca, Mg, K, P, pH in water, and texture. Relative elevation, slope, and cotton yield were also determined. Descriptive statistics, Pearson correlation coefficients, Principal component- stepwise regression and mixed models were used to analyze the data. Soil properties and cotton yield varied widely within relatively small fields and the amounts of variability were field-specific. In general, pH was the least variable of the soil chemical properties. There was no single soil chemical property that could be consistently regarded as the most variable. Yield in the North field was related to K, Mg, and P levels and indicated that management zones based on these nutrients would prove beneficial. Management zones in the South field based on the elevation, sand content, pH, and Ca soil test levels would prove successful if the Ca and pH parameters are used to indicate low productivity areas of the field.