The invention provides a named data network mobility switching method using ELM prediction. An ELM extreme learning machine prediction algorithm is introduced to predict a moving position of a producer, and a consumer can still enjoy seamless content transmission services in the case of switch of the producer by means of a pre-configured network. In the method, an original forwarding table in the network architecture is reconstructed, and control packets are designed so as to realize the pre-configuration of the network, a router is set to store or restore the pre-configuration of the network for response modes of control packets of different types and according to success or failure of a prediction result, and by adoption of the method, the mobility switch of the producer is transparent to consumers. Since a prediction process is introduced, the method is different from an ex post adjustment strategy of the existing mobility switching method, network device pre-configuration is synchronously performed with content transmission prior to the switch, the user response speed is accelerated at a low bandwidth cost, and the additional delay of content response caused by the mobility switch is reduced.