The reasonable scheduling of automated guided vehicle (AGV) plays an important role in the operation of automated container terminal (ACT). It determines the operation efficiency of ACT. In the actual operation process of AGV, the existence of uncertain factors will directly affect the overall operation efficiency of ACT. To respond the uncertain interference in the operation process in time, a digital twin-based AGV scheduling method for ACT is proposed by combining digital twin (DT) with AGV operation mode. DT is the digital expression of physical entity, which has the ability of virtual real interaction and real-time mapping. Through the data interaction between physical space and virtual space, accurate AGV scheduling can be realized. In this paper, DT is applied to an ACT, and DT-based AGV scheduling framework and problem model are established. A DT-based AGV scheduling strategy for ACTs is proposed to further expound the effectiveness and feasibility of the framework and the research problem.