This study develops a multi-objective risk-based optimization framework for Iran’s ‎foreign trade using an integrated approach based on the gravity model, the epsilon-‎constraint method, and a genetic algorithm. Using gravity-model estimations and 2018–‎‎2023 time-series data, the trade potential of 40 major trading partners was evaluated ‎through OLS regression in EViews, and the eight countries with the highest strategic ‎and trade potential were selected. The optimization model was formulated for three ‎commodity groups—energy, petrochemicals, and metals—where export revenue and ‎country risk were defined as the objective functions. The model was solved through a ‎hybrid epsilon-constraint and genetic algorithm approach in the Python environment.‎‏ ‏The results of 10 independent runs of the genetic algorithm indicated satisfactory ‎solution stability, with a coefficient of variation of only 0.60% for export revenue. The ‎optimal solution yielded export revenue of USD 525,208.47, a risk-weighted export ‎volume of 103.22 thousand tons, and a revenue-to-risk ratio of USD 5,087.79 per unit ‎of risk. China, India, the United Arab Emirates, and Türkiye received the largest shares ‎of the optimized export allocation, while the allocation across commodity groups and ‎destinations was shaped by destination-specific risk and commodity characteristics. ‎Relative to the baseline scenario, a 30% reduction in risk increased revenue by 51.83%, ‎whereas a 30% increase in risk reduced revenue by 40.38%. Overall, the findings ‎suggest that export diversification policies should be commodity-specific and designed ‎to account for the trade-offs among export revenue, destination risk, and geographic ‎concentration.‎