Chronic neuroinflammation driven by microglial activation is a pathological hallmark of neurodegenerative diseases, and the NF-$\kappa$B/iNOS signaling axis plays a central role in propagating this damage. NF-$\kappa$B-mediated iNOS transcription generates excessive nitric oxide, causing oxidative neuronal injury. The medicinal mushroom Hericium erinaceus produces cyathane diterpenoid erinacines and isoindolinone erinacerins, both reported to attenuate neuroinflammation; however, the molecular basis of their interactions with iNOS and NF-$\kappa$B remains poorly characterized. We screened 21 erinacerins and 18 erinacines against both targets using validated molecular docking, then subjected top-ranked candidates and negative controls to 100 ns molecular dynamics simulations, MM-PBSA binding free energy calculations (±SEM), per-residue energy decomposition, backbone RMSD, and ligand–protein minimum distance analyses, with quercetin as reference. The analysis revealed scaffold-dependent target selectivity: erinacerins exhibited preferential stability with iNOS (erinacerin L: RMSD 0.185 nm), whereas erinacines formed more stable complexes with NF-$\kappa$B (erinacines G and J: RMSD < 0.36 nm). Minimum-distance monitoring confirmed that the elevated ligand RMSD in iNOS predominantly reflected surface relocation rather than dissociation. Erinacine S emerged as the most promising dual-target candidate ($\Delta G_{bind}$: −24.31 ± 0.16 and −14.24 ± 0.11 kcal/mol for iNOS and NF-$\kappa$B, respectively), over twofold stronger than quercetin for iNOS. Negative controls revealed that docking-based ranking was target-dependent in its discriminative capacity, underscoring the need for MD-based refinement. These results identify erinacine S as a priority candidate for experimental validation.