Reliable and effective medium- to long-term power demand forecasting serves as a crucial foundation for power generation and transmission.With the rapid development of China's renewable energy sector,the impact of wind and solar power volatility cannot be overlooked.Consequently,ensuring that future power system planning can economically and efficiently adapt to varying demand scenarios has become a topic of high concern.Here,we propose an integrated evaluation model for predictive dispatch based on the Extreme Learning Machine (ELM) optimized by the Bat Algorithm (BA),alongside the introduction of fuzzy parameters in the cooperative source-load-storage operation algorithm.Moreover,an analysis and research study has been conducted in northwest China as an example.The results show that this model can accurately forecast power demand under diverse development scenarios and provides scientific guidance for optimizing the planning of source-load-storage resources.