A Mongolian-Chinese machine translation method based on a neural network Turing machine comprises the following steps: firstly, preprocessing Mongolian-Chinese bilingual corpus, vectorizing the Mongolian-Chinese bilingual corpus, and constructing a bilingual dictionary on the basis of the Mongolian-Chinese bilingual corpus; secondly, further expanding storage through a neural network totem machine(NTM), expanding from an internal memory unit of the LSTM to an external memory, introducing a memory mechanism, realizing semantic relation extraction, and giving a semantic relation between two entity words; and finally, searching an optimal solution through decoder model training. Compared with the prior art, according to the invention, semantic analysis is carried out by means of a neural totem machine; related semantic knowledge is found out and extracted, the accuracy of natural language processing is greatly improved by means of the semantic knowledge, corpora are preprocessed by meansof parallel work of a CPU and a GPU, the speed is increased by nearly one time, and the overall translation quality is further improved.