Nickel-Titanium (NiTi) shape memory alloys (SMAs) have been broadly employed in medicine, watch, and automotive industries thanks to their remarkable mechanical properties mainly realized by reversible martensitic transformation. The theory named Phenomenological Theory of Martensite Crystallography (PTMC) is frequently applied to explain the experimental observations obtained by Transmission Electron Microscopy (TEM). Our current understanding of NiTi alloys is however still limited by: a) the characterization of martensite on mesoscale (1 nm - 100 um) which is the dimensional range where the crystallographic interplays with the mechanics by the formation of complex martensitic microstructures is missing, and b) PTMC is phenomenological and does not permit to understand the transformation mechanisms. Therefore, it is significant to revisit the NiTi and other shape memory alloys with new characterization methods such as electron backscatter diffraction (EBSD) and Transmission Kikuchi Diffraction (TKD) techniques and compare the results with the PTMC and another alternative theory called Correspondence Theory. Two types of NiTi alloys will be studied: a martensitic NiTi (with shape memory properties) and an austenitic NiTi (with superelastic properties). The microstructure evolution of the NiTi alloys under deformation will be investigated by EBSD and TKD. The aim of this PhD thesis is to understand the reorientation of the martensitic variants, the twinning modes in martensite, and the texture evolutions in martensitic NiTi, and the martensitic transformation, variant selection and deformation twinning in superelastic NiTi. The interaction work (IW) associated with habit plane variants (from PTMC) or with individual variants (from Correspondence Theory) will be used to predict the variant reorientation or variant selection. A systematic comparison between the two theories will be made. A clear understanding of these events will help us to further explore the nature of the martensitic transformation of NiTi alloys, and promotes their applications by improving the predictions of the microstructure evolution and mechanical properties.