Aortic valve (AV) disease expressed by tissue inflammation and calcification is one of the major causes for cardiac related illness and deaths in the world. Calcific stenosis leads to leaflet stiffening and abnormal AV mechanical-structural performance: larger deformation gradients and stresses. Current AV computational models are somewhat qualitative and do not capture the local complex nonlinear tissue and structural behaviors. Nonlinear multiscale (heterogeneous) material and structural (MMS) models have been developed and used in the analysis of many traditional and advanced engineering materials. These MMS modeling approaches can also be used for the analysis of biomaterials and bio-systems. New refined MMS models for the AV tissue will be developed at the microscale that can be integrated for the mechanical analysis of the AV structure at the macroscale. The overall goal is to introduce refined MMS for predictive simulations of native, porcine and a class of prosthetic AVs under in vitro pulsatile flow conditions. The proposed predictive modeling framework will be verified using sophisticated imagery measurements to examine the kinematics and deformations of both normal and stiffened (diseased) porcine AV systems. A major aspect of this proposal is to introduce new heterogeneous material models for the prosthetic and native leaflet tissue to be used in the AV-MMS simulations. The new modeling approach employs collagen fiber network (CFN) discrete multiscale model depicting the actual bundles of collagen fibers embedded in the leaflet. The in-situ elastin and collagen mechanical properties are explicitly recognized and a unit-cell (UC) micromechanical modeling approach is used to back-calculate (inverse problem) their properties from traditional tissue tensile tests. The proposed MMS is an important research tool towards imagery-based diagnostics of diseased native AV. This paradigm will also allow better design of new prosthetic AV systems.