As modern warfare evolves towards intelligence, the demand for intelligence in weapons and equipment is also increasing. In the military field, equipment intelligence includes intelligent control of equipment, intelligent perception of the battlefield situation, and processing and analysis of command orders, etc., in which technologies such as reinforcement learning and deep learning play an important role. Large Language Model (LLM), as a deep learning model, has a powerful natural language processing capability. Applying LLM to equipment control can improve the autonomy of the weapon system; applying LLM to situation awareness can improve the flexibility and accuracy of situation processing; applying LLM to command analysis can improve the efficiency of command processing. In this paper, after presenting the current status of the application of LLM in weapons research, we analyse the problems that may be encountered. Then, the great potential of the application of LLM is verified by attempting to construct a large-scale command analysis model. Finally, the future trend of the application of LLM in weapons research is predicted.