Knowledge of the spatial variability of soil physical properties is important for precision agriculture. The spatial variation of soil bulk density (r_b), cone index (CI), and gravimetric water content (θ_v) in a loamy sand soil was assessed using several geostatistical techniques and traditional statistics. Field data from a research farm located near Williston, ND was used to evaluate the variability of r_b, CI, and θ_v and to determine whether they were correlated with each other. Soil bulk density was determined from samples collected using a core soil sampler while cone index was measured by inserting a hand-held digital penetrometer into the soil at three different locations within a 30-cm radius of where the r_b samples were extracted. The measurements were made on a 16 m Χ 36 m grid sampling system, which created 72 individual grid cells. Soil properties were measured at the center of each grid cell at depths of 5-10 and 20-25 cm. The three soil parameters were modeled as normally distributed random variables. Sample coefficients of variation were smallest for r_b and largest for CI; a significant correlation was identified between r_b and CI at both depths. Further, a weak correlation was found between CI and θ_v. Sample semivariograms were also constructed for each soil property and were described by combinations of pure nugget and spherical model structures.