At present, research on the refrigerant ratio optimization based on the dynamic boil-off gas (BOG) is limited. Therefore, a new method of dynamic optimization for C3MR liquefaction is presented. Firstly, the simulation program is established based on the steady state model for main liquefaction equipment. Besides, the rules of energy consumption under the range of BOG generation for propane compressor and mixed refrigerant compressor is obtained, and a multi-objective optimization model, which has minimum energy consumption and high heat efficiency, is constructed. Afterwards, the genetic algorithm (GA) and particle swarm optimization (PSO) are designed to determine the optimal results. Finally, the energy efficiency of dynamic BOG is obviously higher than that of steady condition. And compared with the initial simulation process, the maximum energy saving is 50.60%. This study would provide a theoretical basis for the dynamic liquefaction process control and dynamic liquefaction system optimization.