<p>Fine grains are desired because they enhance the overall mechanical performance of alloys through the strengthening effect of grain boundaries. Nevertheless, it is a tough challenge to design parameter loading paths that facilitate continuous deformation toward grain refinement. This work aimed to design variable-parameter loading paths for Ni-Cr-Al alloy to achieve grain refinement during thermal deformation. A three-dimensional continuous processing map was developed to reveal the workability under varying parameters. The effects of processing parameters on microstructural evolution were clarified by combining microstructure characterization with power dissipation efficiency map. The critical power dissipation efficiencies corresponding to different dominant microstructural evolution mechanisms were then identified. A three-dimensional continuous deformation mechanism map, which was derived from processing map, was developed based on these critical power dissipation efficiencies. A three-dimensional continuous parameter domain characterized by stable deformation and dominant dynamic recrystallization (DRX) was separated via the superposition of processing map and deformation mechanism map. Nevertheless, the separated domain was still large and chaotic. To address this issue, the separated domain was further discretized into sub-domains composed of certain values for temperature, strain, and strain rate. Finite element simulations of step-by-step isothermal compression were performed to identify the optimum sub-domains for achieving the finest grain size and largest volume fraction of DRX. Finally, a three-dimensional optimum parameter loading surface was developed by fitting these optimum sub-domains. Experimental results demonstrated that the paths derived from the surface effectively refined grain sizes. This approach offers a novel and generalizable strategy for designing adaptive loading paths in alloys controlled by DRX.</p>