Block yard scheduling is a crucial link in ship construction. To address the issue of low efficiency in yard hoisting operations, this paper constructs a digital twin system and a high-fidelity virtual simulation environment for block yards, and proposes an improved two-stage optimization method named A*HNSA. Driven by the digital twin, this method jointly optimizes the hoisting sequence and the gantry crane path. Based on the real scheduling data from a shipyard in Shanghai, three types of test cases (small, medium, and large) are set up according to the number of scheduling blocks, and comparisons are made with existing methods. The results show that, compared with existing methods in the three types of test cases, the proposed method reduces the total transportation distance by 3.51%, 5.33%, and 6.79% respectively, and shortens the total scheduling time by 3.82%, 5.35%, and 7.07% respectively. It significantly reduces the transportation cost of gantry cranes, improves the solution efficiency, and meets the online response requirements of the block yard digital twin system.