Modern wireless telecommunication devices (GSM Mobile system) (cellular telephones and wireless modems on laptop computers) have the potential to interfere with implantable medical devices/prostheses and cause possible malfunction. An implant of resonant dimensions within a homogeneous dielectric lossy sphere can enhance local values of SAR (the specific absorption rate). Also antenna radiation pattern and other characteristics are significantly altered by the presence of the composite dielectric entities such as the human body. Besides, the current safety limits do not take into account the possible effect of hot spots arising from metallic implants resonant at mobile phone frequencies. Although considerable attention has been given to study and measurement of scattering from a dielectric sphere, no rigorous treatment using electromagnetic theory has been given to the implanted dielectric spherical head/cylindrical body. This thesis aims to deal with the scattering of a plane electromagnetic wave from a perfectly conducting or dielectric spherical/cylindrical implant of electrically small radius (of resonant length), embedded eccentrically into a dielectric spherical head model. The method of dyadic Green's function (DGF) for spherical vector wave functions is used. Analytical expressions for the scattered fields of both cylindrical and spherical implants as well as layered spherical head and cylindrical torso models are obtained separately in different chapters. The whole structure is assumed to be uniform along the propagation direction. To further check the accuracy of the proposed solution, the numerical results from the analytical expressions computed for the problem of implanted head/body are compared with the numerical results from the Finite-Difference Time-Domain (FDTD) method using the EMU-FDTD Electromagnetic simulator. Good agreement is observed between the numerical results based on the proposed method and the FDTD numerical technique. This research presents a new approach, away from simulation work, to the study of exact computation of EM fields in biological systems. Its salient characteristics are its simplicity, the saving in memory and CPU computational time and speed.