A full-parameter identification method based on time-domain mapping neural network of BUCK converter is proposed, which uses the corresponding relationships between output voltage/inductor current and parameters to identify all component parameters. Firstly, a neural network based on the instantaneous relationship is established to identify dominant parameters. Secondly, since recessive parameters are insignificant in the transient relation, an additional network based on the average relationship is built to solve them. The method based on time-domain mapping has a clear physical connection, which contributes to high identification accuracy.