Filtration is a part of the wider concept of contamination control. Contaminants in the hydraulic oil can be in the form of solids, liquids, and gas. Particularly, water in oil can cause major problems in a given hydraulic system such as corrosion and wear of components which would not only deteriorate the efficiency of the system but also lead to extensive maintenance. The type of material used in order to filter water out of the hydraulic oil has to be considered. A filter media is made up of cellulose fibrous media with numerous pores. Due to its property of affinity towards water, the majority of water molecules gets absorbed by its fibrous elements and allowing the oil to pass through it. The filtration efficiency is a function of fluid flow, temperature, operation, and calibration of feed systems. In this thesis a numerical model to estimate the pressure drop over a porous domain is developed. The model is validated using the standard multipass experiment. The CFD model is further developed with 3D random fiber distribution in the porous zone that is used in separating the water in hydraulic oil flows. The developed fibers are distributed over a domain which is equivalent to the overall dimension of the filter media. Ansys fluent’s Discrete Phase Model is used in order to study the water absorption in filter media. The resulting simulation is validated with the water absorption test experiment. The pressure drop and efficiency of the filter media estimated can be used for studying filters properties, type, layer consideration and fiber orientation required for different industrial water absorption applications.