The transition toward a low-carbon and circular economy has increased the demand for sustainable alternatives to petroleum-based packaging materials. Natural fiber reinforced polymer composites (NFRPCs) have gained attention due to their low density, high specific strength, renewability, and biodegradability. However, manufacturing complex three-dimensional packaging structures remains challenging because of fiber heterogeneity, weak fiber–matrix adhesion, and nonlinear thermo-mechanical interactions during processing. This study proposes an integrated framework combining multi-scale mechanics modeling, experimental validation, and multidisciplinary design optimization (MDO) for the development of high-performance green packaging components. Mechanical behavior was predicted using the rule of mixtures and Halpin–Tsai models, while Darcy’s law and heat transfer equations were employed to analyze resin impregnation and thermal evolution. A real-time filament winding and weaving platform was established to obtain experimental data. Box–Behnken design (BBD) and Response Surface Method (RDM) were used to evaluate the effects of heating temperature, winding tension, winding speed, and roller pressure on tensile strength, porosity, and energy consumption. The Non-dominated Sorting Genetic Algorithm (NSGA-II) was then applied to identify Pareto-optimal process parameters. Experimental validation confirmed the accuracy of the optimized results, the results obtained using the optimized process show that there is a 10.46% improvement in tensile strength, 37.40% decrease in porosity, and 1.00% decrease in specific energy consumption as compared to the original process, showing the practical value of the suggested NSGA-II optimization approach, providing a reliable approach for designing sustainable packaging materials from agricultural waste.