Spatial variability of soil properties such as pH and pH buffering capacity cannot be accurately measured by traditional soil sampling and laboratory analysis, because these properties are highly variable in the field, requiring large numbers of samples. Therefore, spatial characterization of these soil properties for routine use is cost prohibitive. Measurement with an appropriate sensor would therefore be useful if it can be done with relatively high accuracy and low cost. Our objectives were to determine if Near Infrared Reflectance Spectroscopy (NIR) could be used to estimate these soil properties with a laboratory grade instrument and to determine if accurate sensing in the field might be possible with a reduced wavelength range instrument. Fifteen to 25 soil samples were taken at a depth of 7.5 cm from each of five fields in South Georgia. Sampling locations were selected visually to cover the range of soil organic C and clay concentrations in the fields. Samples were dried at 30 C and ground to pass a 2 mm sieve. The soils were analyzed by measuring soil pH in 0.01 M CaCl2 (1:1 soil to solution), and measuring soil pH buffering capacity with a titration procedure. The soils were scanned from 400 to 2500 nm (Si and PbS detector) using ISI quarter cups and transport module in a computer controlled NIRS system, model 6500 scanning monochrometer (Foss-NIR Systems, Silver Spring, MD), with data collected every 2 nm and using the Win ISI II, version 1.5 software. NIR reflectance measurements, R, were transformed to log 1/R and subsequently first derivatives were calculated. The spectra were also analyzed using only a portion of each spectrum, from 950 to 1650 nm. Modified partial least square (MPLS) regression techniques were used to develop the calibration equations. First, all data was used for calibration followed by prediction of the measured values using the calibration data set. Then, the calibration data set used two thirds of the data with every third data point left out for validation. When all data was used for the calibration of pH and pH buffering capacity, the resulting predicted pH and pH buffering capacities were highly related to the values measured in the laboratory (R2 = 0.97 for pH and 0.77 for pH buffering capacity). When two thirds of the data were used for calibration to predict the remaining one third, the regression of predicted vs lab measured values yielded R2 values that ranged from 0.91 to 0.95 for pH and from 0.53 to 0.69 for pH buffering capacity. When the spectra was truncated to 950 to 1650 nm and all data was used for calibration, the resulting predicted pH and pH buffering capacities were still well related to the values measured in the laboratory (R2 = 0.81 for pH and 0.69 for pH buffering capacity). When two thirds of the data were used for calibration to predict the remaining one third, the regression of predicted vs lab measured values yielded R2 values that ranged from 0.52 to 0.74 for pH and from 0.38 to 0.64 for pH buffering capacity. When the predicted soil pH and pH buffering capacity were used to calculate lime recommendations, predicted and measured values agreed well. Using the truncated spectra and calibrating with two thirds of the data, R2 values ranged from 0.35 to 0.67 and Root Mean Square Errors ranged from 300 to 600 kg lime per ha on data with a range of lime recommendations from 60 to 3500 kg lime per hectare. These results are promising for use of NIR technology in the field.