Shape memory alloys (SMAs) are one of the most popular smart materials, with the ability to release residual strain and return to their initial configuration when heated up to a particular temperature. Due to phase transformations during thermomechanical loading, the constitutive modeling of these materials is complex. In most of the existing three-dimensional phenomenological models, some of the material parameters cannot be directly evaluated and must be calibrated by comparing the obtained responses with experimental results. This dissertation concerns the modeling of shape memory alloys using microplane formulation in a thermodynamically-consistent framework. Microplane formulation provides closed form relations for calculating the strain components in terms of the stress components. One advantage of this approach is in the limited number of the required material parameters. These material parameters can be calculated in simple tension and torsion tests. The main feature of this modeling is in describing the behavior of a complex material with simple constitutive laws on each microplane. To this end, the approach finds the material behavior in any direction on each microplane and then uses the micro-macro homogenization process to obtain the overall macroscopic properties. Traditionally, microplane formulation was based on Volumetric-Deviatoric-Tangential split and macroscopic strain tensor was derived using the principle of complementary virtual work. It has been shown that this formulation might violate the second law of thermodynamics in some loading conditions. Here, a free energy potential is defined at the microplane level. Integrating this potential over all orientations provides the macroscopic free energy. Based on this free energy, a new formulation based on Volumetric-Deviatoric split is proposed. These formulations in a thermodynamic framework capture the behavior of shape memory alloys that will maintain the second law of thermodynamic. The new formulation based on V-D split is compared numerically with V-D-T split in uniaxial and pure torsion. In order to extract material parameters, experimental works are essential. So, the superelastic response of NiTi shape memory alloys under various loading conditions is experimentally studied using thin-walled tube specimens. In addition to proportional loading, several non-proportional loading experiments based on both stress/strain control modes are conducted. Furthermore, several combined tension-torsion experiments are conducted, and the coupling between tension and torsion in NiTi tubes is investigated. The proposed model is validated with experimental data. This model is able to simulate both the shape memory effect and superelasticity. Deviation from normality is experimentally studied in nonproportional loading and is compared with proposed model. Also, cyclic behavior and asymmetric in tension and comparison is investigated by proposed model and are compared with experimental work. The good agreement seen between the results indicates the reasonable accuracy of the proposed microplane approach in studying the SMA behaviors at different conditions. In continue; propose model is extended to capture anisotropic behavior in SMAs. Coupling between tension and torsion is another issue that is experimentally studied and is compared with microplane model. A numerical procedure is developed to implement the proposed model as a user material subroutine (UMAT) in the nonlinear finite element software ABAQUS/Standard commercial code. Several real case such as NiTi spring and Retrograde blade are simulated based on the proposed model. Regarding these results, proposed model can be used to simulate various SMA applications. Keywords : Shape memory alloys; Microplane model; Thermodynamically-consistent framework; Volumetric-Deviatoric split; Multiaxial loading; Experimental work.