Expansive soils, which have been reported as a worldwide problem, cover 25% of the United States. Due to the swelling and shrinkage behavior induced by moisture variations, expansive soil contributes to volumetric deformation, which in turn affects the stability and performance of structures. The Texas Department of Transportation (TxDOT) allocates 25% of its budget to pavement maintenance and repairs, much of which is triggered by expansive soil. In order to decrease the burden of this expense on maintenance authorities, it is necessary to have an accurate understanding of expansive subgrade behavior. Applying this knowledge to the pavement design and construction processes can significantly increase the pavement's service life. The specific objectives of this research were to (1) study the behavior of expansive soil with seasonal changes and climatic loading; (2) asses the real-time moisture and temperature variations in the expansive subgrade; (3) quantify the deformation pattern with time in response to environmental loading; (4) develop a realtime moisture, temperature, and deformation prediction model; (5) based on the investigation of the subgrade, provide solutions in order to combat the pavement deformation; and (6) evaluate the effectiveness of the proposed solution. In order to accomplish the objectives, one farm-to-market road and one state highway were selected for observation of the behavior of expansive subgrades in North Texas. Soil samples were collected and tested to determine the soil properties. Moisture, suction sensors, temperature sensors, and rain gauges were installed to record the variations of the variables over time. Moreover, geophysical testing was conducted to continually portray the subgrade over time. Deformation of the pavement was monitored through topographic surveying and a horizontal inclinometer. Collected data was analyzed in a statistical environment to develop real-time prediction models. The first attempt produced a moisture variation model that captured variations due to seasonal effects and temporary variations due to rainfall. The outputs of this model were within 90% of the values measured on-site. The second attempt produced a temperature prediction model that was dependent on depth and the day of the year. The squared correlation coefficient between the observed and predicted soil temperature was more than 0.90. Application of the developed models could allow for a non-invasive estimation of the response of soil strength and stiffness properties due to variations in moisture and temperature. While examining the deformation data, it was found that seasonal variations only capture a portion of the deformation, whereas the amount of precipitation plays a significant role in further modifying the model. Temperature and suction were also correlated with deformation to finalize the deformation model. Application of the developed model facilitates estimation of deformation at any time of the year, in response to precipitation. The study also attempted to focus, to a limited extent, on numerical modeling; however, the selection of unsaturated parameters was challenging. The selection of unsaturated permeability and flow parameters is usually laboratory-based, because a specific condition of the soil makes it impossible to capture them in real time in the field. This study attempted to determine the variations of unsaturated hydraulic conductivity based on rainfall response data. Rather than conducting the usual laboratory testing to determine the unsaturated flow parameters by curve fitting, a novel approach was undertaken to determine the flow parameters from field soil water characteristic curves. Finally, field-based values were used in the PLAXIS 2D environment for transient analysis. The validity of the estimated parameters was confirmed, as FE results corresponded with direct field measurements. The study results indicated that FE modeling can provide effective information about the subgrade matric suction variations. This research focused on finding a possible solution to the problem of pavement distress. It was found that controlling the moisture from the edge of the pavement can significantly improve the pavement performance. Consequently, a moisture barrier consisting of a geomembrane and a geocomposite (geonet sandwiched between two nonwoven geotextiles) was suggested. A combination of a 40-mil LLDPE geomembrane and an 8-oz. HDPE geocomposite was used to control the moisture from the edge of a 50 ft. section of FM 987. A control section along the same roadway was instrumented and monitored for comparison. Preliminary field monitoring results clearly indicated that the moisture barrier significantly reduced the water infiltration near the edge of the pavement. Moreover, the movement of the pavement was reduced by 80%, based upon previous recorded measurements of the control section.