Damage identification in mechanical systems under vibration excitation relates to the monitoring of the changes in the dynamical properties of the corresponding linear system, and thus reflects changes in modal parameters (frequencies, damping, mode shapes) and finally in the finite element model of the structure. Damage localization can be performed using ambient vibration data collected from sensors in the reference and possibly damaged state and information from a finite element model (FEM). Two approaches are considered in this framework, the Stochastic Dynamic Damage Location Vector (SDDLV) approach and the Subspace Fitting (SF) approach. Both localization algorithms are presented and their performance is illustrated and compared on simulated and experimental vibration data.