Current intelligent agent decision control systems lack comprehensive self-assessment of operational status and effective decision-driving mechanisms, often resulting in limited adaptability to emergencies and unknown environments, as well as constrained task performance. To address the demands of intelligent agent decision-making, this paper proposes a novel brain-inspired decision-making method based on Spiking Neural Networks (SNN). First, a spiking encoding method for multi-source heterogeneous data is introduced to transform the agent’s multi-source state data into spiking signals. Subsequently, leveraging the encoded multi-source state information, a brain-inspired decision control computational model based on SNN is established to achieve real-time, accurate, and stable state identification and decision-making for the agent. The proposed brain-inspired intelligent decision-making method provides a technical solution for developing next-generation intelligent agent decision systems, with the potential to accelerate the advancement of intelligent agents toward greater autonomy and intelligence.