The macroscopic simulation of soot production in flames by CFD approaches often assumes of spherical particles or considers ad-hoc formulas in order to take the particle fractal aggregates’ morphology into account. On the other hand, numerically simulated aggregates can be generated by Discrete Element Modeling (DEM) simulations at the nanoscopic scale. However, the change in thermodynamic conditions, surface growth, oxidation, and nucleation mechanisms are commonly neglected by this approach. This work combines these complementary approaches to investigate the detailed morphology of soot along 4 different particle trajectories in a diffusion flame. The proposed multi-scale approach shows remarkably larger and more compact aggregates near the wings of the flame as compared to the centerline, in agreement with previous experimental studies. This approach allows to analyse the particle polydispersity and morphological parameters beyond the fractal dimension, such as anisotropy coefficient and monomers overlapping coefficient. It becomes possible to reveal clear morphological signatures of soot formed along different streamlines in the flame. Finally, we have also observed a positive correlation between primary particle diameter and aggregate gyration diameter due to the predominance of surface reactions over aggregation. The proposed post-processing of CFD results based on DEM methods, bringing new information on soot morphology, is a proof of concept of a more accurate procedure for validating soot models used in CFD codes as compared to existing methods in the literature.