Previous studies by the authors have shown that if properly implemented, the Pennes assumptions can be applied to quantify bioheat transfer during extremity heating. Given its relative numerical simplicity and its ability to predict temperatures in thermoregulated tissue, the Pennes model of bioheat transfer was utilized in a three-dimensional thermal model of limb heating. While the arterial blood temperature was assumed to be radially uniform within a cross section of the limb, axial gradients in the arterial and venous blood temperatures were computed with this three-dimensional model. A realistically shaped, three-dimensional finite element model of a tumor-bearing human lower leg was constructed and was "attached" mathematically to the whole body thermal model of man described in previous studies by the authors. The central as well as local thermoregulatory feedback control mechanisms which determine blood perfusion to the various tissues and rate of evaporation by sweating were input into the limb model. In addition, the temperature of the arterial blood which feeds into the most proximal section of the lower leg was computed by the whole body thermal model. The variations in the shape of the tissues which comprise the limb were obtained from computerized tomography scans. Axial variations in the energy deposition patterns along the length of the limb exposed to a miniannular phased array (MAPA) applicator were also input into this model of limb heating. Results indicate that proper positioning of the limb relative to the MAPA is a significant factor in determining the effectiveness of the treatment. A patient-specific hyperthermia protocol can be designed using this coupled electromagnetic and thermal model.