Energy utilization and environment protection has received increasing attention over the last few decades. Therefore, flare minimization has become one of the main concerns for the ethylene industry. In this study, we present an improved flare minimization model for ethylene plant start-up, which accounts for different types of cracking furnaces and production/consumption of high-pressure steam. The resource-task network approach is used to depict the superstructure of the start-up process of an ethylene plant. On the basis of the superstructure, we develop a mathematic formulation, which, in this study, is a mixed-integer linear programming model. Five instances of an industrial case study with different start-up working medium and high-pressure steam inventories are used to demonstrate the efficacy of the proposed flare minimization model.