The increasing complexity of biopharmaceutical production systems and the growing emphasis on sustainability necessitate the development of intelligent, data-driven decision-making frameworks capable of handling multiple conflicting objectives under uncertainty. This study proposes a data-driven multi-objective optimization model for sustainable biopharmaceutical production planning by integrating Artificial Neural Networks (ANN), Genetic Algorithm (GA), Particle Swarm Optimization (PSO), the ε-Constraint method, and Monte Carlo Simulation. In the first phase, an ANN-based model was employed to forecast product demand using real industrial data, achieving a promising accuracy with RMSE ≈ 869. The forecasted demand was then embedded into a multi-objective mathematical model designed to maximize economic profit while minimizing environmental impacts. The model was solved using GA and PSO, and the Pareto-efficient solutions were generated through the ε-Constraint technique to visualize trade-offs between objectives. Results indicated that PSO outperformed GA in both convergence speed and solution quality, particularly for large-scale and nonlinear decision spaces. Furthermore, Monte Carlo simulation was utilized to perform uncertainty and sensitivity analysis, confirming the robustness of the proposed model against fluctuations in demand and environmental parameters. The Pareto frontier revealed a dynamic and controllable trade-off between profitability and ecological sustainability. Overall, the integration of machine learning, metaheuristic optimization and probabilistic analysis provides a comprehensive intelligent framework for data-driven, sustainable and resilient production planning in the biopharmaceutical industry.