In this study, the water content and dry density data from field measurement of the ATC Engineered Roadway Soil (E-RW) were analyzed first, and their means and standard deviations were derived. Based on corresponding ERDC lab test data, each soil parameter in both material model and EOS equation was then expressed as a function of soil water content and dry density by 3D or 4D surface fitting. Thus mapping equations from soil water content and dry density to soil parameters were established. With an assumption of normal distribution for both soil water content and dry density, a stochastic soil model was developed for the ATC Engineered Roadway Soil. Modeling and simulation examples were provided to demonstrate how to apply the developed stochastic soil model to carry out underbody blast Monte Carlo simulation for generating vehicle and occupant responses clouds, and how to estimate the low and up bounds of occupant and structure response with certain confidence level.