Cardiovascular system diseases (CVDs) are the leading cause of death in developed and developing countries. One of the most frequent group are valvular diseases, caused among others by the calcification or congenital defects, resulting in stenosis and many severe medical complications. CVDs present a significant burden for the healthcare system. Thus, there is a room for enhancement for diagnostic and therapeutic procedures, which will lead to treatment improvement and costs reduction. One of the remedies are the computer tools, which are used for the virtual therapies and support the medical diagnoses. Presented work describe the development and procedure for modeling the human aortic valve using two-way partitioned fluid-structure interaction algorithm. The two-dimensional geometry used in the model includes the valve, aortic root and adjacent aortic and ventricular artery portions. The advanced moving mesh model consists the coupling of the dynamic mesh smoothing and the overset mesh technique, to speed up the computation and improve the convergence and stability. This model was compared with the simpler dynamic mesh procedure. The pulsable inlet boundary condition based on the measurement data was implemented. For the outlet boundary condition in the limited cardiac system geometry, the Windkessel lumped parameter model has been included within the solving procedure. The calculations were performed for the various severity of the atherosclerosis, resulting in different hemodynamic characteristics. The multifluid Euler-Euler approach with the interaction between the phases described by the kinetic theory of granular flow (KTGF) was compared with the single phase model. The literature data were partially used to check the feasibility of the developed model.