Additive manufacturing (AM) of metal alloys enables fabrication and repair of complex parts and with lesser material wastage. However, the qualification of AM components for structural applications remains a challenge due to strong processing based variability in mechanical behaviour of AM alloys. Alloy properties such as yield strength, strain hardening, and fatigue resistance are governed by microstructures which in turn depend strongly on thermal histories. These microstructures exhibit significant variations in grain morphology, crystallographic orientation, and phase distribution, which controls the local stress concentration and plastic strain localization. Therefore, a reliable prediction of mechanical properties and performance requires micromechanics-based understanding of how processing conditions influence microstructures and deformation. In this work, we present an integrated experimental and computational framework to quantify microstructure-sensitive mechanical properties in additively manufactured Ti-6Al-4V produced using Selective Laser Melting. Printed coupons are subjected to post-build heat treatments to transform the as-built martensitic structure into an α+β microstructure. Microstructures are characterized using scanning electron microscopy and electron backscatter diffraction, and their mechanical response is evaluated through hardness and tensile testing. We then generate statistically equivalent synthetic microstructures based on characterized descriptors of grain size, morphology, and crystallographic orientation. These synthetic microstructures are analysed using crystal plasticity finite element simulations (CPFEM) calibrated against experimental measurements to predict macroscopic stress–strain behaviour and local strain localization patterns. This methodology enables a systematic assessment of the role of microstructural variability on yield behaviour and deformation localization. Plastic strain hotspots are critical precursors to damage initiation and fatigue. A reduced-order surrogate model is also developed to rapidly map microstructural features to mechanical properties, to identify damage-critical regions. The integrated framework provides a mechanics-based pathway for linking AM processing routes to microstructure-controlled mechanical reliability.