Process parameters in Laser Powder Bed Fusion (L-PBF) significantly influence the thermo-mechanical performance of fabricated parts. This effect is especially critical for shape memory alloys (SMAs), where processing conditions can alter microstructural characteristics, leading to changes in transformation temperatures and, consequently, functional performance. Accurate thermal modeling of SMAs during the L-PBF process is therefore essential to control process-induced transformations and defects. In this work, we propose a temperature-dependent laser absorption coefficient specifically tailored for NiTi alloys under L-PBF conditions. A mesoscopic thermofluid model is developed using computational fluid dynamics (CFD) to simulate melt pool behavior at the powder scale. The temperature-dependent absorptivity is determined by adjusting the model to reduce the difference between simulated melt pool sizes and measurements taken from high-resolution images of individual laser tracks. This inverse calibration approach yields a physics-informed absorption profile that accurately reflects the thermal response of NiTi during processing. The resulting model shows significantly improved agreement with experimental melt pool geometries, enhancing the reliability of thermal predictions. This work provides a foundation for simulation-driven process control and alloy design in the additive manufacturing of functional SMAs.