Abstract: This study addresses the critical challenge of layer density optimization in Multi-Layer Insulation （MLI） systems operating under specific structural constraints, a key factor for minimizing heat ingress in cryogenic applications. An enhanced inter-layer thermal radiation model, accounting for variable shield emissivity, is introduced. Genetic Algorithm （GA） and Particle Swarm Optimization （PSO） are rigorously employed as complementary optimization tools. This research systematically optimizes MLI structures utilizing diverse metal foil coatings （aluminum, gold, copper） as radiation shields and explores novel performance enhancement pathways under combined conditions of variable shield coatings and strategically designed non-uniform density distributions. The GA was first applied to analyze the intricate relationship between coating configurations and optimal layer density variation patterns. Simulation results demonstrate conclusively that the hybrid multi-metal coating structure achieves a significant 33.54% improvement in overall insulation performance compared to conventional single-metal-coated MLI. Crucially, configurations where layer density decreases monotonically from the cold end to the hot end exhibit substantially lower total heat leakage than those employing an increasing density gradient. Quantitatively, the total heat flux is measured at 0.3903 W/m² for the decreasing-density configuration versus 0.4109 W/m² for the increasing-density case, unequivocally validating the efficacy of directionally graded density design based on fundamental heat transfer principles. Building on this, the PSO algorithm was subsequently deployed to optimize individual layer spacing parameters within a segmented three-zone structure （low-density, medium-density, and high-density zones）. This advanced optimization procedure yielded the globally optimal inter-layer spacing combination for the partitioned system. The resulting configuration achieves a remarkably low minimum total heat leakage of 0.3916 W/m², representing a distinct 2.27% enhancement over the performance of a comparable uniform-density baseline or a non-optimized zoned structure. This work contributes robust, algorithm-driven optimization methodologies and delivers novel insights into synergistic multi-metal coating strategies combined with zoned density architectures. These findings provide a practical and significant framework for substantially advancing the thermal performance of next-generation, high-efficiency MLI systems in demanding thermal environments.