The filter material in embankment dams is a crucial part of the dam due to protecting the core material (usually clay) against internal erosion. Internal erosion is defined as the transportation of core material particles to the filter material by seepage flow. It is the main reason for the deformation and loss of resistance in the body and foundation of a dam. Therefore, it is a serious threat risk to embankment dams. To prevent internal erosion, it is necessary to accurately evaluate the interaction between water-filter material-core material during the dam's design and operation stages. A suitable filter material should be able to control and block cracks that can form in the core material. Terzaghi (1925) did the first research on filter material and proposed a criterion based on the particle size of a clay. Another detailed study was conducted by Sherard et al. (1989) who developed an experiment called No Erosion Filter Test (NEF Test) to comprehend the relationship between filter and core materials. NEF Test is seen as the most proper method to determine the critical filter which leans on the impermeable core material in the downstream side of an embankment dam. In the NEF test, the most severe condition that can occur is represented as an erosive leak from the core material to the filter material. Within the scope of this study, the internal erosion of the backfill used in the dam is studied through outflow rate and deformations in the base sample (clay). First, the samples taken from two embankment dams in Izmir (Kalabak Dam and Rahman Dam) were brought to the geotechnical laboratory of Izmir Institute of Technology (IZTECH), and the filter material was mixed with fine materials (silt) in different contents. A series of NEF Tests were performed in the IZTECH-geotechnical laboratory on the filter and core samples of Kalabak Dam within a TUBITAK Project No. 221M071 (Ecemis, 2023). Then, the performed NEF Tests were modeled by the finite difference method (FDM) using the Itasca-FLAC3D software. Finally, the outcomes of the numerical models were compared to experimental results. Numerical models had a good agreement to decide whether filter sand performance is successful or not.