Abstract: New energy units equipped with energy storage have the potential to serve as blackstart power sources. Studying their black start capability during the initial black start stage is significant for building an emergency defense support system suitable for the new power system. Considering the spatiotemporal variation characteristics of the black start capability of new energy stations, a ternary table containing elements such as load level, source load electrical distance, and start service safety level is proposed to characterize the black start capability of new energy stations quantitatively. Based on this, define the black start space-time support capability. The short - and long-term memory neural network and knowledge map are introduced, in which the short - and long-term memory neural network is used to predict the security level of startup services quickly and then assess the space-time support ability of black startup. The knowledge graph stores and visualizes the rolling updated data. Finally, the future application scenarios for the rolling evaluation of the black start spatiotemporal support capacity of new energy stations were illustrated with examples.