Abstract: In response to the challenges faced by the gas field group in the eastern South China Sea during the production process, such as the difficulties in collaborative production, significant fluctuations in output, and the difficulty in optimizing production strategies, the objective is to develop an integrated and intelligent digital twin platform to achieve real-time knowledge, rapid simulation, and intelligent decision-making of the production system. Based on the CS/BS hybrid architecture, a three-layer software platform comprising the data layer, service layer and application layer was designed and developed. On the basis of this architecture, machine learning algorithms have been deeply integrated to construct a high-precision and high-efficiency integrated digital twin model for the gas reservoirs, wellbores and pipeline networks. The engineering implementation was carried out by combining Python and .NET Core, and the automatic historical fitting and real-time correction of model parameters were achieved by using dynamic production data. Field applications show that the platform has successfully achieved real-time and seamless integration with on-site production data. It can automatically correct model parameters and complete minute-level full system state prediction. The trial indicates that the speed of physical process simulation calculation is improved by more than 200 times compared to traditional methods, significantly shortening the research and adjustment cycle of geological oil reservoirs, and supporting real-time calculation of the optimal production system. This achievement has effectively addressed the technical challenges faced by gas field group in collaborative production and output optimization, significantly enhancing development efficiency, production management level, and scientific decision-making capabilities.