Automated container terminal (ACT) plays an important role in the modern logistics and transportation industry. The utilization of automated guided vehicles (AGV) can be effectively promoted by reducing their empty running. The existing scheduling strategies of AGVs for most ACTs cannot ensure their full load, leading to underutilization. This work proposes a double-cycling AGV path planning model to maximize the loading rate of AGVs by scheduling their loading and unloading tasks. The randomness of the quay crane’s operational time is considered. By allocating time intervals of AGVs arriving at the quayside, it determines a loading sequence of containers based on a genetic algorithm with a penalty function. Experimental results show that the model can realize double-cycling transport with AGVs and effectively reduce the operational cost of ACTs.