The diffusion of digital technology has brought a subversive impact on the technological development of various industries, but the diffusion path and formation mechanism of digital technology in various industries have not been clear. In this paper, the patent data of 97 Chinese national economic industries from 2000 to 2019 were used to construct the citation network, and the Search Path Count (SPC) and critical path algorithm were used to identify the main path of digital technology diffusion among industries. This paper analyzes the formation mechanism of the main path of digital technology diffusion from four dimensions of technology characteristics by using exponential random graph model (ERGM). Through the identification of the diffusion main paths, it is found that the 13 digital technology diffusion main paths are disconnected and involve significant differences in technology fields, core technologies, major cities and institutions, possibly due to the influence of industry boundaries and industry heterogeneity. The results of the exponential random graph model show that technology combinatorial opportunity and technology proximity have significant positive effects on the formation of diffusion main path, while technical knowledge diversity and technical cooperative potential have significant negative effects on the formation of diffusion main path. In addition, the endogenous network structure variables have a significant positive impact on the formation of the main path of digital technology diffusion. These findings reveal the current trend and direction of China's digital technology development, which provides an important basis for technology developers and policy makers to choose the path of digital technology development, and provides effective guidance for traditional industries to grasp the development opportunities brought by digital technology.