A method for predicting plant lncRNA-miRNA interaction, comprising: obtaining an original data set, processing it to obtain a set of lncRNA sequence features, a set of interaction profiles of lncRNA, a set of miRNA sequence features, a set of interaction profiles of miRNA, and calculating The similarities S_lncS, S_lncP, S_miS, and S_miP corresponding to the lncRNA sequence feature set, the lncRNA interaction profile set, the miRNA sequence feature set, and the miRNA interaction profile set respectively, the sequence feature similarity of lncRNA S_lncS, and the interaction profile of lncRNA The similarity S_lncP is combined to obtain the combined similarity of lncRNA, and the sequence feature similarity S_miS of miRNA and the interaction spectrum similarity S_miP of miRNA are combined to obtain the combined similarity of miRNA; from the first interaction matrix 1 / 5 of the elements are randomly selected and set to zero to obtain the first training matrix.