The authors discussed a Newton-type iterative method for solving a nonlinear ill-posed inverse electromagnetic scattering problem: the quantitative reconstruction of the complex permittivity distribution of inhomogeneous 2-D or 3-D objects from scattered near-field data. The use of an iterative procedure is important because the effects of ill-conditioning can be significantly reduced by enforcing the convergence with information about object external shape, upper and lower bounds of complex permittivity, the presence of different media, etc. The domain of validity of this method is directly related to the signal-to-noise ratio. The method is illustrated with various numerical simulations of practical interest performed on dissipative inhomogeneous dielectric objects (biological phantoms, human arm, etc.) in view of biomedical applications of microwave imaging. The influence of noise has been also studied in order to test the robustness of the algorithm, and reconstructions using real experimental data have been made