Aiming at the problem that the recognition accuracy ofsingle neural network model decreases significantly with the increase of the number of emitters in individual emitter recognition,this paper proposes an ADS-B emitter individual recognition method based on deep neural network and random forest ensemble model.This method uses a variety of deep neural network models and random forest to train the enhanced data set,and then uses votingclassifier to ensemble the results obtained by each network model and random forest recognition,which makes the recognition results more convincing.The experimental results show that,after intagrating the models of DRSN,VGG,ResNet,GoogleNet, DenseNet neural networks and random forest,the recognition accuracy ofensemble model can be improved by 3%～20% compared with the single neural network,and the recognition accuracy can still be maintained with the increase of the number of individual radiator.