The large language model, exemplified by ChatGPT, has brought a disruptive impact to the field of artificial intelligence, with a primary focus on natural language processing, speech recognition, machine learning, and natural language understanding. This paper innovatively applies the large language model to the field of intelligent decision-making, places the large language model in the decision-making center, and constructs an Agent architecture with the large language model as the core. Building on this, it further introduces a two-tier Agent task planning strategy, issuing and executing decision commands through natural language interactions, and conducting simulation validations within a wargame simulation environment. Through game confrontation simulation experiments, we find that the intelligent decision-making capability of large language models is significantly superior to that of commonly used reinforcement learning AI. This superiority is apparent in terms of intelligence, comprehensibility, and generalizability. And through experiments, it is found that the intelligence of the large language model is closely related to Prompt engineering. This work also expands the application of large language models from previous human-computer interactions to the realm of intelligent decision-making, providing valuable insights and significance for the advancement of intelligent decision-making.