The identification of genes responsible for human diseases based on functional consistency and network topological features is of great importance for both understanding human disease pathogenesis and improving clinical practice. A novel method based on the functional consistency and network topological features was introduced to establish an association between genes and diseases. Using this method, candidate disease genes were predicted from disease risk loci. Then, the candidate genes sharing the same or similar functions with known disease genes in the functional enrichment analysis of GeneOntology and KEGG databases as final disease genes were determined. 51 genes were predicted to be the disease genes for coronary artery disease and most of them participate in the development of disease by literature retrieval. The method provided additional insights for the finding disease genes, which will be helpful for the studies on the pathogenesis of human complex diseases.