Soot aerosol generated from incomplete combustion of biomasses, fossil fuels and biofuels is a major light-absorber, however its spectral complex refractive index (m=n-ik) remains uncertain. In this study, fractal soot aerosols with varying maturity and physico-chemical properties were studied systematically in a large simulation chamber to determine their complex refractive index. The real (n) and imaginary (k) parts of m were retrieved through optical calculations using the measured size, morphology, density and particles’ scattering, absorption and extinction coefficients as input parameters. Optical calculations were performed using Mie theory, which represents soot as homogeneous spheres, and the Rayleigh–Debye–Gans for fractal aggregate (RDG-FA) theory, which accounts for soot aggregate morphology. Monte Carlo simulations were used to quantify how uncertainties in input parameters affect the retrievals. Results from the analysis show that both Mie and RDG-FA theories can reproduce the measured absorption and scattering of soot, but that the retrieved n and k values differ substantially between the two approaches, underscoring the need for theory-consistent use of refractive index values in optical modelling. For soot with elemental-to-total carbon ratios of 0.79 (BC-dominated) to 0 (BrC-dominated), retrieved values span 1.35–2.68 (n) and 0.38–0.31 (k) using Mie theory, and 1.68–2.07 (n) and 0.20–0.09 (k) using RDG-FA. Mie-based refractive indices are weakly sensitive to uncertainties in input parameters but strongly affected by particle coagulation and size growth, reflecting limitations of the spherical approximation. The RDG-FA complex refractive index retrievals remain stable during coagulation but depend strongly on the uncertainty of the assumed particle morphology.