To address the technical bottlenecks such as large data volume and limited transmission bandwidth encountered in constructing a high-fidelity digital twin of a liquefied natural gas (LNG) carrier, this paper proposes a ship-to-shore data transmission system that integrates reliable transmission and efficient compression. First, the system uses standardized interfaces to achieve multi-source data acquisition, and through a data hierarchical transmission strategy, extracts the value of the data. On this basis, the GZip algorithm is introduced to efficiently compress key data such as those from the liquid cargo tank system. Second, based on the data transmission architecture of the Message Queuing Telemetry Transport (MQTT) protocol, a quality of service (QoS) level confirmation mechanism is introduced to ensure effective data transmission, and an intelligent network disconnection recovery strategy is proposed to enable local caching at the ship end and ensure complete data transmission. Finally, real ship application verification shows that the proposed system achieves a compression rate of over 93% while ensuring 100% complete data transmission, and significantly reduces the end-to-end transmission delay under limited bandwidth. This study comprehensively improves the effectiveness of data transmission for the digital twin of LNG carriers from three dimensions—transmission efficiency, reliability, and real-time performance—and verifies the effectiveness of the proposed transmission strategy and modules, laying the foundation for the engineering application of high-value ship digital twin systems.