In the last two decades, the use of modal parameters to detect and localize structural damage has gained increasing interest in the scientific community. The work reported in the literature mainly deals with the change of natural frequencies due to changes in structural properties given by damage. This approach relies on the accurate modal parameters estimation, which is still a very difficult task especially when complex structural systems are of interest. Furthermore, damage detection and localization can be highly inaccurate if natural frequencies are used only, regardless of the associated mode shapes. This paper investigates the effectiveness of vibration-based methods for dynamic identification and damage detection on masonry structures using experimental tests and operational modal analysis. In particular, ambient vibration tests are first used to obtain modal parameters of two different building models made up of unreinforced and confined masonry, respectively. A shaking table is then used to give a sequence of increasing seismic loads to each of the two masonry models. Damage-induced variations of the modal parameters, i.e., natural frequencies and associated mode shapes, are estimated after each seismic level using the structural vibrations caused by a low-level white noise base excitation. The obtained results give useful information on the effectiveness of vibration-based methods to detect and localize different damage patterns on masonry structures using changes in the modal parameters. Furthermore, it was possible to assess the different seismic performance of the confined masonry structural system compared to the unreinforced masonry.