Managing spare parts inventory for aircraft is one of the more persistent challenges in aviation maintenance, sitting at the intersection of operational continuity and environmental responsibility. Airlines must prevent costly parts shortages that could ground aircraft while simultaneously containing the financial burden and carbon footprint associated with excess stock. Existing approaches tend to treat these objectives separately, leaving efficiency gains on the table. This research proposes a multi-objective optimisation framework for sustainable aviation supply chains that combines machine learning forecasting, circular economy principles, and environmentally conscious logistics planning. The framework comprises three integrated components: a feature engineering pipeline augmented with sustainability indicators, a hybrid XGBoost-LSTM demand prediction model incorporating sustainability-weighted loss functions, and a data-driven circular economy recommendation engine that classifies parts by reuse, repair, or disposal potential. Validation on an industry-calibrated dataset shows an 18.7% reduction in total annual costs, a 23.4% decrease in carbon footprint, and service levels maintained above 93%. These results indicate that operational efficiency and environmental accountability are complementary rather than competing objectives in aviation MRO contexts.

Aircraft spare parts inventory management represents one of aviation maintenance's most enduring challenges, where operational continuity meets environmental responsibility. Airlines face the difficult task of avoiding expensive parts shortages that can ground aircraft while also controlling the financial costs and carbon emissions that come with holding too much inventory. Current methods typically address these goals in isolation, missing opportunities for greater efficiency. This study introduces a multi-objective optimization framework for sustainable aviation supply chains that integrates machine learning forecasting, circular economy concepts, and environmentally aware logistics planning. The framework consists of three connected elements: a feature engineering process enhanced with sustainability metrics, a combined XGBoost-LSTM demand forecasting model that uses sustainability-weighted loss functions, and a data-driven circular economy recommendation system that categorizes parts based on their potential for reuse, repair, or disposal. Testing with an industry-standard dataset demonstrates an 18.7% reduction in total yearly costs, a 23.4% drop in carbon footprint, and service levels sustained above 93%. These findings suggest that operational efficiency and environmental responsibility work together rather than against each other in aviation maintenance, repair, and overhaul operations.