This application provides a method for predicting a node state, including: obtaining static graphs and dynamic graphs of a plurality of nodes in a target network, where the static graphs and the dynamic graphs are all topology views; generating spatial feature data of the plurality of nodes based on the static graphs and the dynamic graphs; obtaining time feature data of the plurality of nodes; and obtaining a predicted state of a target node in a target time range based on the spatial feature data and the time feature data, where the target node is any node in the plurality of nodes. The method for predicting a node state provided in this application is applied to the field of node state prediction in a network, and uses a dynamic spatial feature in addition to a time feature and a static spatial feature. This increases accuracy of node state prediction in a network with a dynamic topological relationship, so that the network is more intelligent in node state prediction.