The knowledge graph has helped the development of knowledge application and neural net-
work subject with its intricate ，information- intensive and semantically related characteristics. The de-
velopment of knowledge representation learming methods and the increasing demand for cognitive intel-
ligence have made knowledge representation and reasoning a research focus in the current knowledge
graph and natural language processing fields. Large-scale knowledge reasoning based on representa-
tion learing has also been verified on relevant data sets. This article mainly explores the following
work:①The existing representation learming and reasoning methods are sorted out, and the knowl-
edge graph reasoning methods are divided into : Distance model method，semantic similarity method of
bilinear model, neural network learming method ，graph Neural network learning method.②The arti-
cle investigates the prospective future of the representation learning and reasoning methods of the
knowledge graph. The development of representation learning and reasoning methods of the knowledge
graph is of great significance to the entire knowledge community and to help the development of strong
artificial intelligence.