<p>We propose a new mechanistic framework to unveil the fundamental mechanisms governing multi-cycle plastic strain recovery in nanocrystalline metals. The model uniquely integrates crystal plasticity in nanograins with grain boundary (GB) chemo-mechanics, explicitly resolving atomic flux driven by chemical potential gradients under evolving stress and free volume distributions. Applied to nanocrystalline copper films, our simulations capture transient (10⁻⁷ s⁻¹) and steady-state (10⁻⁸ s⁻¹) strain recovery rates spanning hours to days, achieving quantitative agreement with experimental kinetics across six orders of time scale. Three key advances emerge: (1) GB-mediated atomic diffusion dominates recovery (contributing &gt; 75% of total strain reversal), while dislocation back-stress in nanograins plays a secondary role; (2) recovery cycles induce microstructural evolution through stress-driven free volume redistribution, generating chaotic GB stress states and localized plasticity accumulation at triple junctions; (3) macroscopic strain recovery masks progressive microplasticity in GB networks, revealing a fatigue precursor mechanism inaccessible to conventional models. This work establishes the first predictive link between atomic-scale GB dynamics and macroscopic time-dependent recovery, providing a transformative tool for designing fatigue-resistant nanocrystalline alloys through GB engineering.</p>