An electroquasistatic (EQS) model of capacitive hyperthermia for treating lung tumors is proposed, based on which the finite element method is applied to compute the electrical potential in a human thorax model. The temperature distribution in the thorax model, which is surrounded by a bolus maintained at a constant temperature, is computed by numerically solving a bio-heat equation, which includes metabolic heat generated in the tissues, heat convection mechanism in tissues and bolus, as well as the heat delivered by the microwave field computed with the EQS model and finite element method. Temperature-dependent blood perfusion rates of blood and muscle, respectively, are adopted to account for the physiological reaction of tissues to temperature variation. By simulations, it is observed that adjusting the dielectric properties of adipose tissue via injection, the time evolution of temperature distribution can be controlled to some extent, providing more flexibility to customize a hyperthermia treatment plan for specific patient.