This article presents the concept of the digital twin of a city. The urban economy is a complex multi-vector system. Creating a single digital twin of such a system is now a difficult task to solve. In this article the concept of a digital twin of a city is presented. The authors propose an evolutionary approach to this problem, according to which digital twins of individual elements of the urban environment are consistently built on a single hardware and software platform. These digital twins are linked in a single cooperative system that allows one digital twin to use data produced by other digital twins. The article gives the definition and architecture of such a system. It describes the classes of models that can be used to create digital twins. Special attention is paid to neural network models and models for data analysis. The information infrastructure of the digital twin of a city, including sensory networks, data cleansing mechanisms and nebulous calculations, is considered.