Inefficient irrigation and the excessive use of water on agricultural lands in the Aral Sea Basin over several decades have led to highly saline soils. Salinity appraisal in the Aral Sea Basin however, is still dependent upon traditional soil surveys with subsequent laboratory analyses for the Total Dissolved Solids (TDS). This study has three specific objectives, namely, to identify techniques that enable rapid estimation of salinity, to characterize the spatial distribution of soil salinity and to estimate the spatial distribution of soil salinity based on readily or cheaply obtainable environmental parameters. Soil salinity was measured by four Electrical Conductivity (EC) devices (2XP, 2P, 4P, and CM-138) on a regular grid covering an area of approximately 3 km by 4 km. Six nested samplings within selected grids were conducted to account for small-scale variation. The farm-scale (~15 km²) results were used to upscale soil salinity to a district area (~400 km²). Apart from widely used terrain indices and those acquired from remote sensing, distance to drains and long-term groundwater observation data were used to account for local parameters possibly influencing soil salinity. Standard statistical procedures were applied for data description, correlation between variables, analysis of variance, and regression. Characterization of the spatial distribution of soil salinity and interpolation of point data were carried out using geostatistics. Soil salinity estimation based on environmental attributes was carried out using a neural network model, as this offers enhanced generalization compared to other models. Analyses were integrated into a GIS for visualization and presentation of the results. Techniques for rapid determination of soil salinity based on electrical conductivity were assessed and proved to be satisfactory in all cases. Measurement based on soil paste (EC_p) was highly accurate (R²=0.76), whereas EC_a measurements at point scale in the field were of low accuracy (R²<0.5). However, field assessment of soil salinity was considerably enhanced by the use of CM-138, because large areas can be quickly assessed, which in spite of lower accuracy, is desirable. Topsoil (30 cm) salinity was highly variable even at short distances (40 m) compared to average soil salinity at 0.75 m and 1.5 m depth measured by the CM-138. Overall distribution of soil salinity was influenced by soil texture and topography, while at the local scale terrain attributes such as curvature, plan and profile curvatures, and solar radiation were the most influential factors. Factors obtained by remote sensing had significant correlation coefficients (r=0.2) with both the salinity of topsoil and salinity measured by the CM-138. Distance to drains is an important factor, especially for the bulk soil salinity (measured by the CM-138) of the profile. Correlation between distance to drains and salinity of the topsoil was low, which might be due to higher spatial variation of the topsoil salinity. Groundwater table depth and salinity had marked correlations with soil salinity; however, the direction of the influence could not be explained. The inclusion of these controlling variables in modeling is fundamental, and efforts must be directed towards obtaining reliable and accurate databases in order to derive them. With an environmental correlation model that was built for the farm scale, soil salinity was estimated using environmental parameters in a neural network approach and shows a high correlation coefficient between estimated and measured soil salinity of 0.83. The accuracy of the prediction of soil salinity was satisfactory taking into account that the measurement scales of soil salinity and environmental data derived from different estimations with unknown but certainly varying accuracy. The use of environmental attributes and soil salinity relationships to upscale the spatial distribution of soil salinity from farm to district scale resulted in the estimation of essentially similar mean soil salinity values (0.94 dS m^-1 vs. 1.04 dS m^-1). However, visual comparison of the maps suggests that the estimated map had soil salinity that was overly uniform in distribution, which is thought to be caused by inaccuracy of environmental data (including scale problems) or overgeneralization by the neural network model.