<p>Digital transformation has long been widely applied across industries through ICT, and digital twin technology as a leading innovation, is further driving the digital upgrade of complex industrial systems. Due to the complexity of structure hierarchy, data interaction and system connectivity in the complex industrial systems, there are still challenges in the application of digital twins, such as difficulties in model design, complex data management, and high software platform development costs (as developer costs, license fees, and outsourcing costs). This paper proposes a solution that consists of the model design for digital twins and a corresponding software architecture platform for complex industrial systems. Firstly, a six-dimensional model for digital twins and a four-dimensional modeling theory for virtual entities is proposed to solves the problem of model design for digital twins; Then, a dual-core data architecture, which combines relational databases and time series databases (called R-TSDB), is proposed to optimize data storage and access for the digital twin model. Finally, a digital twin software platform architecture based on a six-dimensional model that integrates multiple model types and handles multi-source data, with community-driven initiatives supporting the reduction of implementation costs. Extensive experiments conducted on a 600MW subcritical coal-fired boiler at a power plant demonstrate that the proposed solution is of great significance to the modeling and implementation of digital twins in complex industrial systems.</p>