Traditional procurement demand forecasting methods primarily rely on historical data and static models, making it challenging to respond in real-time and effectively to sudden events and rapidly changing market dynamics. This paper proposes a procurement demand forecasting method based on the dynamic fusion of multi-source data, which deeply integrates expert decision-making systems and web crawlers’ technology. The main tasks include: 1. Utilizing web crawlers and API to dynamically obtain real-time information on earthquakes, weather, holidays, etc., while combining pre-set strategies to comprehensively assess the current emergency situation or specific scenario. 2. When external information triggers procurement forecasting, the system performs a detailed segmentation of demand, combining the materials database and Global Product Classification (GPC) standards to intelligently recommend the specific names, categories, priority levels and recommended rationale. 3. The system is built by using PyQt6, enables hourly automatic detection, seven-day procurement forecasts, querying past events, automatic report generation, and procurement strategy adjustments. Test results under various emergency scenarios show that the system achieves instantaneous response in all types of emergency scenarios, and the procurement recommendation results are highly consistent with actual needs.