Atomic reconstruction reshapes moiré materials across multiple scales, from local structure and polarization textures to global electronic topology, yet direct ab initio modeling becomes prohibitive for large superstructures such as marginal-twist-angle moirés and moiré-of-moirés. We develop MoiréMLIP by fine-tuning a universal atomistic model on the density functional theory labeled Moiré Kaleidoscope dataset, which spans transition metal dichalcogenide compositions, symmetries, stackings, and twist angles. MoiréMLIP reproduces ab initio reconstruction with force errors of 6–8 meV/Å and transfers to smaller twist angles and unseen structures. Knowledge distillation yields MoiréMLIP-mini, which retains this accuracy while extending single-GPU inference to one million atoms. Applied to an alternate-twist MoTe₂ trilayer, it reveals a hierarchical polarization network spanning tens of nanometers arising from large moiré-cell distortions sharply localized along the moiré-of-moiré domain walls. These results overcome key accuracy, transferability, and scaling bottlenecks of existing atomistic models and enable predictive simulations across emergent moiré length scales.