By-product gas scheduling during iron and steel production is crucial to saving energy and reducing emission. However, the dynamic fluctuations of gas supply and demand, pipeline network pressure and gas cabinet capacity as well as uncertainties caused by production conditions changes often expose the disadvantages of existing scheduling models. Therefore, a multi-objective synergistic dynamic scheduling model for by-product gas is proposed, which integrates economic model predictive control (EMPC) and an improved adaptive differential evolution with optional external archive (IJADE) optimization to achieve synergistic optimization of the production system and by-product gas system. A by-product gas scheduling model is constructed by minimizing an objective function with purchased electricity cost, gas dissipation and equipment adjustment and considering constraints with balance between supply and demand of gas and power generation efficiency of unit. The model can respond quickly to changes of working conditions by using a dynamic adjustment mechanism. The global search capability and convergence speed of JADE are enhanced by using a rolling time-domain dimension reduction, an initial population improving approach with hot-startup and terminal constraints. The experimental results show that compared with JADE, the proportion of self-generated peak hours has been improved by 1.1–1.7%; the cost of purchasing electricity has been reduced by 50,000 RMB/day; the gas discharge rate and the mean square error have been decreased.