The work aims to investigate the multi-depot packaging waste vehicle routing problem considering fuzzy emission to study the impact of uncertain packaging waste generation on collection routes and improve the quality of collection operations. First, based on the fuzzy credibility theory, triangular fuzzy numbers were introduced to represent waste generation at collection points, and an optimization model was constructed with the objective of minimizing collection mileage while incorporating the decision maker's subjective preference constraints. Secondly, an Adaptive Large Neighborhood Search (ALNS) algorithm with 12 neighborhood criteria was designed. Finally, in the case study, different waste collection schemes under both deterministic and fuzzy emission conditions were compared, employing various intelligent optimization algorithms and conducting sensitivity analysis on the decision maker's subjective preference values. The collection mileage considering fuzzy emission was 262.34 km longer than that under deterministic demand. The ALNS algorithm outperformed Ant Colony Optimization (ACO) and three other algorithms in terms of solution performance, with the optimal subjective preference value for the decision maker being 0.6. This study effectively addresses the challenges posed by uncertain emissions, providing theoretical support and reference for related fields.