Abstract Digital twin technology simulates physical systems through real-time data and virtual models, providing an innovative solution to challenges in power communication systems, such as multiple access interference (MAI) and the near-far effect. By constructing a digital twin model, we can accurately measure the non-orthogonality of each user’s spread spectrum code, enabling the elimination of mutual interference through matrix inversion or iterative methods. Our simulations show that the application of digital twin technology improves system performance by reducing MAI and stabilizing communication quality under complex conditions. When the number of stages increases, the system performance improves, but when it is greater than level 5, the system performance does not improve significantly, so the level 5 is selected in the practical application. In this case, the interference factor of the grid communication system is equal to 0.55, and the resulting system gain is 2.