Shale is often the most difficult of all formations to maintain a stable wellbore in when drillincr ::> for oil and gas. Time and money spent overcoming this problem during drilling, together with overall reduced profit margins. has led the oil industry to devote considerable time and effort to solve the problem of unstable boreholes in shales. It has long been established that the moisture adsorption (or desorption) of shale rocks can be controlled by the salinity of drilling fluid. When compacted shale (under constant compaction stress) adsorbs moisture, its total volume increases and swelling strains develop. Developed swelling strains then become an integral part of the effective radial stress acting on the shale formation contributing to borehole failure. A mathematical model has been developed for predicting the swelling behaviour of shale when placed in contact with water under moderate pressures and the effect of the swelling on borehole (in)stability. The model is based on thermodynamic theory which suggests that fluid movement into or out of a shale is driven by an imbalance in the partial molar free energy of the shale and the contacting fluid. Conversion of the free energy of each system (fluid and shale) into "total swelling pressure" made it possible to model transient pressures and strains generated in shale. The analytical solution of the radial diffusivity equation is reduced to a simpler form for the model. The model was validated using equipment and experimental techniques which allow continuous monitoring of shale swelling as function of time and distance from the wetting end. It was found that increasing the compaction stress acting on the shale reduced the rate of swelling, and increasing the hydraulic pressure of the fluid on the shale's wetted surface increased the rate of swelling. This behaviour was adequately described by the model which therefore represents a new method for predicting shale swelling as function of time and radial distance under different environments. Swelling strains are then used to predict related changes in shale mechanical properties (failure criteria) and well (in)stability. Several well-site index tests have been developed to study shale-drilling fluid interaction at wellsite. These index tests can provide input data for the mathematical model. Drilling fluids can be screened for their ability to control shale swelling, thus minimising the risk of well bore instability.