The invention relates to a graph clustering method based on perception application algorithm semantics and a computer readable medium. The graph clustering method includes: acquiring original graph data G and a graph's application algorithm A; initializing a subgraph Gi; randomly selecting a vertex v and a corresponding connecting edge thereof from the graph G, and deleting the v and the corresponding edge thereof from the graph G; computing a semantic serial degree after adding the vertex v into the subgraph Gi; determining a clustering block with the maximum semantic serial degree, and adding the vertex v and the corresponding connecting edge into the subgraph; repeating the steps until the graph G is empty; and completing graph clustering, and outputting a clustering result. Compared with the prior art, the method provided by the invention has the advantages of being able to greatly accelerate application analysis and mining of big graph data.