<p>NiTi shape memory alloys are widely used in biomedical and aerospace applications, yet their fibre-laser weldability remains constrained by thermal sensitivity, intermetallic formation, and a persistent trade-off between tensile strength (TS) and corrosion resistance (CR). This study addresses that gap by integrating nanopowder-assisted fibre-laser welding with a stacked-surrogate NSGA-II optimisation framework. Three nanopowder fillers (Fe, Cu, Ni) were employed both as microstructural modifiers of the weld zone and as a discrete design variable within the optimisation. A 43-run central composite design varied laser power, pulse time, frequency, wire diameter, and filler type, with TS and CR measured as simultaneous quality indicators and fracture surfaces analysed by SEM-EDS. The stacking ensemble achieved RMSE values of 17.75&#xa0;MPa for TS and 1.8 × 10⁻⁴&#xa0;mm/yr for CR, both within the repeatability range of the corresponding experimental measurements, confirming their practical acceptability. The NSGA-II Pareto front revealed three application-relevant regimes: Cu-filled welds offer the lowest corrosion rates suited to long-term biomedical implant environments; Fe-filled welds deliver the highest strengths for load-bearing applications; and Ni-filled welds provide a balanced trade-off appropriate for aerospace actuators where both fatigue and corrosion resistance are required. SEM-EDS evidence supported these regimes mechanistically. The framework provides a transferable, pre-fabrication toolkit for selecting fillers and tuning process parameters in high-performance NiTi joints, reducing reliance on trial-and-error development.</p>