<p>Additive Manufacturing (AM) is widely regarded for its disruptive potential, presenting an opportunity for promising reductions in environmental impact. Current sustainability optimization approaches in AM are based on optimizing “indirect” factors such as energy consumption, printing time, quality of the additively manufactured parts, etc. However, such approaches fall short in offering a holistic understanding of AM’s sustainability. Thus, there is a pressing need for a more seamless integration between assessment and optimization tools for AM’s sustainability. Further, there exists a strong demand for accurate quantification and early design stage optimization of AM’s environmental impacts, directly using their respective values. In pursuit of these objectives, this study proposes a novel hybrid framework that integrates product-process codesign, Life Cycle Assessment, Machine Learning, and Metaheurtics to develop data-driven sustainability assessment and optimization for AM. Building on this foundation, the study introduces a two-stage optimization framework that supports the development of a multi-criteria sustainability-induced optimization tool. Integrated into the hybrid framework, this tool functions as a feedback mechanism to guide early-stage design modifications for improved sustainability outcomes. The efficacy of the proposed methodologies is validated through a case study on the Material Extrusion process. The Non-dominated Sorting Genetic Algorithm II is employed to solve a multi-objective minimization problem, encompassing various environmental impact categories and average part surface roughness. Results reveal that the algorithm is found effective in solving the multi-objective problem. To identify the most optimal choice from the Pareto solutions, the Order of Preference by Similarity to Ideal Solution approach is employed. The study concludes by discussing its limitations and highlighting future research endeavors.</p>