Abstract: Ethylene distillation column (EDC) is one of key low-temperature and energy-consuming equipment in ethylene plant and its operating status have direct effects on the quality of products. In the past decades, most of studies have been focused on the modeling and simulation of EDC and few works have been reported about how the EDC operates during its startup process. Even in the summary of industrial startup experience, there has rarely been deep analysis on EDC individually. Therefore, this work is aiming at analyzing the dynamic characteristics of EDC and optimizing its startup operation strategy (SUOS). Firstly, both the steady and dynamic model of EDC will be built through Aspen HYSYS. Historical data in industry is then used to verify the simulation results. And then, some more tests will be applied to illustrate that the trend of system response and the range of time-delay are both close to the facts. By adding frequent disturbances from random changes of feed flowrate, the proposed control scheme can achieve a good stability performance. Moreover, in order to attain better reflection of real case, the dynamic modeling is integrated with the SUOS programmed in MATLAB to implement the EDC industrial startup process. According to the industrial specifications, EDC startup procedure must be separated into two parts, i.e., total-reflux operation (TRO) and start-up operation (SUO). TRO and SUO could be privately divided into discontinuous and semi-continuous stage according to dynamic performance of hydraulic and thermodynamic variables. Finally, a time-constrained optimization problem (TCOP), in which the way of adjusting reflux flowrate is connected with energy consumption, will be proposed during TRO. An evolutionary algorithm, termed as the feasibility rule with the incorporation of objective function information (FROFI), is utilized to solve this TCOP problem. Experimental results show that compared with initial process, the proposed method can save 25% energy and could provide effectively theoretical guidance for the energy-saving potential during real EDC startup process.