The growing need to understand the behaviour of un-reinforced masonry URM), subjected to repeated light man-made earthquakes caused by the extraction of gas in the north-eastern part of The Netherlands has resulted in intense research to determine the exact process of crack initiation and propagation. The historical masonry buildings and Dutch terraced houses in Groningen are prone to light damages which become severe upon repeated lateral earthquake loading. Although there are material models that describe the behavior of modern brick masonry, they do not accurately represent the mechanical properties of 19th century clay brick masonry. This led to a large-scale research into the mechanical behavior of un-reinforced masonry and an orthotropic continuum macro-model called the Engineering Masonry Model (EMM) was proposed. The existing tension constitutive model in EMM assumes a secant unloading-reloading branch which does not consider the strength degradation of URM under repeated loading. Since tension mode-I fracture results in cracking of URM, it is important to study the effects of repeated loading on the propagation of the crack and its effects on the capacity of the structure. This thesis presents a degradation model to represent the strength deterioration of URM observed during repeated loading. The constitutive model formulated in this thesis is based on hyperbolic functions along with a secant slope for the unloading-reloading branch. To justify the model assumptions, a single linear 4-node element is analysed with the new model and the effect of varying different components of the constitutive equations is established. The window bank spandrel sample modeled as a 4-point bending test is analysed using the new model for 10, 30 and 100 repetitions. It is shown that the hyperbolic model can predict accurately the stress reduction within each repetition displacement set and also represent the crack width widening and crack propagation accurately when compared to the experimental results. The new model is tested on a wall with a window opening sample and the results closely matched that of the experiment. Finally, recommendations are provided for further development of the hyperbolic model and calibration of the material properties.