The province of Quebec, in Canada, is located in a stable region of the North American tectonic plate. But some zones are seismically active. Moreover, Montreal City has the second highest urban seismic risk in Canada. Although it is now recognized that unreinforced masonry (URM) is vulnerable to earthquakes, many URM buildings are critical infrastructures built before provisions were made for seismic loading. A research project has been undertaken to improve understanding of unreinforced masonry seismic behaviour. It is part of the Canadian Seismic Research Network (CSRN) for the seismic evaluation and rehabilitation of critical infrastructures. This research project, consisting of three main parts, is a reference document to be used in future research work. The first part consists of a review of literature on the seismic behavior of unreinforced masonry, specifically on masonry, the effects of time, and various types of masonry works. Next, special attention is devoted to the description of failure modes. Finally, a review of factors influencing the seismic performance of masonry walls is carried out by using the following four specific properties: stiffness, strength, deformation capacity and energy dissipation capacity.The in-plane and out-of-plane loadings of masonry walls with and without frames are discussed separately. There are some behaviour similarities between a wall without a frame and an infill wall. The second part consists of the experimental evaluation of the in-plane cyclic behavior of a steel frame infilled with structural clay tile masonry wall that is representative of a first floor external back wall of a typical school built in Quebec in the 1950's. Its development is based on structural inventories for vulnerability assessment. Results are analyzed according to stiffness, strength, deformation capacity and energy dissipation capacity of the experimental specimen. Some characterization testing has been done to determine properties of masonry such as compressive strength, modulus of elasticity and Poisson's ratio. Unlike the characterization results, the experimental wall has shown a reserve capacity after masonry cracking. The third part aims to validate the capacity of the SeismoStruct's masonry model to predict the experimental response of infill walls specimens. This model, developed by Crisafulli, is based on the equivalent strut concept often used for infill wall modelisation. Experimental data obtained in this project and a similar project are used to verify whether this model could be used in practice with rational settings for the seismic evaluation of a building.The methodology used did not yield satisfactory predictions.