The research of open innovation (OI) has developed considerably in recent years. In this article, a new perspective based on the patent citation network is provided to explore the dynamic evolution and mode of OI. In our framework of the OI network, enterprises are represented as nodes, and patent citations represent cross-organizational knowledge flow, which are ties in the network. The Driver Assistance System (DAS) was selected as the research case. Time-sliced patent citation networks are constructed, and then an exponential random graph model is employed to identify the formation mechanism of OI networks. The results show that the OI network of DAS is still partially open and at a low level. The inherent dominance of automakers may have been weakened, and new models and relationships in innovation activities are developing. In addition, heterogeneity in type and geographic proximity which significantly promote the formation of the open network was prevented, but the larger scale of the enterprise inhibited the OI network. R&D investment has no obvious impact. This research provides a new perspective to observe open innovation and helps stakeholders to better understand industry trends.