The United States of America has naturally fractured reservoirs containing many tens of billions of barrels of oil. When secondary and enhanced recovery methods are applied, fractures tend to channel injected fluids through the reservoir to production wells, resulting in much of the oil in the matrix blocks being bypassed and not recovered. This results in a low recovery efficiency from fractured reservoirs. Because of the large size of this resource, any new technology that improves recovery efficiency from naturally fractured reservoirs by as little as a few percent could result in a significant increase in overall oil production from existing petroleum formations. Oil recovery by steam injection is a proven, successful technology for nonfractured reservoirs, but has received only limited study for fractured reservoirs. Preliminary studies suggest recovery efficiencies in fractured reservoirs may be increased by as much as 50% with the application of steam relative to that of low temperature processes. The key mechanisms enhancing oil production at high temperature are the differential thermal expansion between oil and the pore volume, and the generation of gases within matrix blocks. Other mechanisms may also contribute to increased production. These mechanisms are relatively independent of oil gravity, making steam injection into naturally fractured reservoirs equally attractive to light and heavy oil deposits. The objectives of this research program are to quantify the amount of oil expelled by these recovery mechanisms and to develop a numerical model for predicting oil recovery in naturally fractured reservoirs during steam injection. The experimental study consists of constructing and operating several apparatuses to isolate each of these mechanisms. The first measures thermal expansion and capillary imbibition rates at relatively low temperature, but for various lithologies and matrix block shapes. The second apparatus measures the same parameters, but at high temperatures and for only one shape. A third experimental apparatus measures the maximum gas saturations that could build up within a matrix block. A fourth apparatus measures thermal conductivity and diffusivity of porous media The numerical study consists of developing transfer functions for oil expulsion from matrix blocks to fractures at high temperatures and incorporating them, along with the energy equation, into a dual porosity thermal reservoir simulator. This simulator can be utilized to make predictions for steam injection processes in naturally-fractured reservoirs. Analytical models for capillary imbibition have also been developed.