Reconfigurable and morphing structures may provide significant improvement in overall platform performance through optimization over broad operating conditions. The realization of this concept requires structures, which can accommodate the large deformations necessary with modest weight and strength penalties. Other studies suggest morphing structures need new materials to realize the benefits that morphing may provide. To help meet this need, we have developed novel composite materials based on specially designed segmented reinforcement and shape memory polymer matrices that provide unique combinations of deformation and stiffness properties. To tailor and optimize the design and fabrication of these materials for particular structural applications, one must understand the envelope of morphing material properties as a function of microstructural architecture and constituent properties. Here we extend our previous simulations of these materials by using 3D models to predict stiffness and deformation properties in variable stiffness segmented composite materials. To understand the effect of various geometry tradeoffs and constituent properties on the elastic stiffness in both the high and low stiffness states, we have performed a trade study using a commercial FEA analysis package. The modulus tensor is constructed and deformation properties are computed from representative volume elements (RVE) in which all (6) basic loading conditions are applied. Our test matrix consisted of four composite RVE geometries modeled using combinations of 5 SMP and 3 reinforcement elastic moduli. Effective composite stiffness and deformation results confirm earlier evidence of the essential performance tradeoffs of reduced stiffness for increasing reversible strain accommodation with especially heavy dependencies on matrix modulus and microstructural architecture. Furthermore, our results show these laminar materials are generally orthotropic and indicate that previous calculations of matrix gap and interlaminar strains based on kinematic approximations are accurate to within 10-20% for many material systems. We compare these models with experimental results for a narrow geometry and material set to show the accuracy of the models as compared to physical materials. Our simulations indicate that improved shape memory polymer materials could enable a composite material that can accommodate ~30% strain with a cold state stiffness of ~30GPa. This would improve the current state of the art 5-10x and significantly reduce the weight and stiffness costs of using a morphing component.