As technology competition between countries and companies becomes increasingly intense, various commercial and policy means are being used to hinder the technology development of competitors. The risk of technology development is thus much higher than before and must be accurately assessed so that timely preventive measures can be taken to ensure the safety of technology development. In this paper, a hybrid data-knowledge driven method is proposed for constructing a technology risk evaluation indicator system that can accurately identify the high-risk technologies. Expert knowledge is effectively leveraged through a data-driven model developed from a novel double-layer bagging machine learning method, which is able to learn the indicator aggregation rules in the indicator system automatically. A case study on the risk assessment of smart chip-related technologies for China is provided. As a result, the high-risk areas such as “field programmable gate array” and “central processing unit” are accurately identified, manifesting the effectiveness and accuracy of the proposed method.