The light and frequent earthquakes in the north of the Netherlands, particularly in the province of Groningen, have recently exposed the unreinforced masonry structures to seismic activities. Since these structures do not adhere to seismic regulations, they are considered vulnerable to seismicity. The ultimate state capacity of the structures is important for an individual's safety, however, these earthquakes are of low intensity and cause aesthetic damage. In order to investigate the light damage initiation and development, TNO has performed shaking table tests on an unreinforced masonry (URM) cavity wall specimen in out-of-plane (OOP) one-way bending with small increments in intensity. The test specimen consisted of calcium silicate brick inner leaf and perforated clay brick outer leaf. The damage development in the outer leaf was monitored during these tests using a high-speed digital image correlation (DIC) technique to study the initiation and development of damage in the outer leaf of the specimen. The experimental tests showed damage initiation at the mid-height of the outer leaf. The tests could not capture the development of cracks through the thickness of the cavity wall. The scope of this research is a numerical assessment of the experimental study by using a Non-Linear Time History (NLTH) analysis of light damage initiation and development of a URM cavity wall under out-of-plane loading. The high-resolution experimental results are used as a basis for the development and calibration of models which can better predict the crack initiation and development in URM. The finite element software DIANA 10.5 FEA was used to set up the numerical model and conduct transient analysis. The seismic signal as an input loading and the top boundary condition of the test specimen. The acceleration data measured from the shaking table tests at the base was used as an input seismic signal for the transient analysis of the models. The input signal needed to be processed before application as the presence of low-frequency content leaded to inaccurate results. Different approaches are discussed in this thesis regarding the processing of the input acceleration signal. The experimental tests were modeled along the cross-section of the test specimen, thereby highlighting the thickness of the inner leaf and the outer leaf. This enabled tracking the light damage initiation and propagation through the thickness of the cavity wall. A total of thirteen shaking table tests were conducted on the experimental setup. In order to gain insight into the behavior of the specimen during each shaking table test, a model was created corresponding to each shaking table test. Preliminary analysis schemes were set in order to check the validity of all thirteen models. The two cases of top boundary conditions were checked, roller support and spring-mass support. The roller boundary condition proved to be stiff in comparison to the experimental results. The numerical results were calibrated on the basis of material properties. The results were compared to experimental results by checking the dynamic behavior at the mid-height, dynamic behavior over the height, and light damage initiation and development of the specimen. The results of the numerical models were stiff in comparison to the experimental results. According to the conclusions, it is recommended to research further regarding the boundary conditions, especially the bottom boundary condition due to the formation of a rocking crack. Another important aspect to focus on is the combination of all input signals, thereby, taking into consideration the damage accumulation.