Abstract: In recent years, the proportion of renewable energy in China's power system has gradually increased, the randomness and uncertainty of the system have become more and more obvious, and the quantitative evaluation of power system flexibility has become the core issue of the safe and stable operation of the system. Therefore, this paper proposes an error distribution calculation method based on historical data screening, which can better characterize the uncertainty of the system based on historical data, combined with Spearman coefficient and non-parametric kernel density estimation method. Secondly, the characteristics of thermal power units, energy storage, and load-side flexible resources are taken into consideration, and an evaluation method for system flexibility supply capacity is proposed. Then, the limitation of the critical section is taken into consideration, and the influence of the limitation on the flexibility of the system is analyzed; moreover, the system flexibility evaluation method considering the transmission limitation of the critical section is proposed. Finally, the modified IEEE 30-bus system is simulated, and the results show that the proposed prediction error probability distribution can truly reflect the uncertainty faced by the system, and provide theoretical supports for flexible quantitative analysis in complex "source-network-charge-storage" scenario.