The international community has been operating several export control regimes to protect human lives and property from weapons of mass destruction. The Korean government also implements export control of strategic items and technologies by reflecting relevant guidelines in the Foreign Trade Act and Nuclear Safety Act. The first step in implementing export control is to determine whether the item is to be classified as a strategic item or not, which is conducted through a process of self-classification by exporters or a classification request to the government. According to relevant Korean laws and regulations, exporters must apply to the government for an item classification and proceed with export licensing if the result corresponds to strategic items. In the field of nuclear technology, it is difficult to secure open data due to the strict security standards imposed by the nature of the industry, and continuous experience and expertise in nuclear technology are required for classification judgment. Therefore, this study analyzes the possibility of using deep-learning technology to determine whether items and technologies are strategic items and to utilize the history of the Korea Atomic Energy Research Institute (KAERI)'s strategic items export and import to find related information and support the prediction of strategic item classification. In addition, with KAERI's data, the simulation was conducted based on a specific word model and a deeplearning model suitable for the development of the strategic item classification system was explored. After preliminary tests, the results show that a few neural network models were found to be effective for the characteristic of documents, and the necessity of pre-processing of the document for accuracy improvement was confirmed. The results of the study will be used to develop a decision support system to determine whether a strategic item belongs to the nuclear trigger list, which will contribute to securing KAERI's strategic trade control system and improving practical efficiency.