This study focuses on an alumina industrial park equipped with a small-scale wind-PV hybrid system, and conducts optimization of production processes to achieve the goals of energy saving, carbon mitigation, and cost reduction. The non-dominated sorting genetic algorithm II (NSGA-II) is employed to derive typical daily energy dispatch configurations. First of all, the alumina production processes using the Bayer method is comprehensively investigated, and optimized improvement schemes tailored to the characteristics of each production link is proposed. Secondly, the impact of link optimization on industrial alumina manufacturing is quantitatively analyzed. Thirdly, NSGA-II algorithm is introduced, and a tri-objective optimization model targeting maximal alumina output, minimal carbon emission, and optimal production expenditure is constructed. Finally, an energy allocation scheme based on the features of seasonal multi-energy generation is developed through combined with synergetic optimization strategy. The study shows that after links optimization and model optimization, the carbon emissions of the alumina industrial park equipped with wind-PV hybrid systems supplying 995 million kWh power annually have been reduced by more than 90%, and the daily alumina output is more than 6,000 tons, ensuring the alumina annual output of 2 million tons is achieved. Meanwhile, considering buying and selling electricity, the production costs of the park will be reduced because of an annual income of 34,000 yuan obtained in the production processes. This study can provide a basis for operational decision-making, so that to achieve the effect of ‘stable output, carbon reduction, and cost savings’.