Abstract: Diabetes is a complex disease caused by the combined effects of genetic and environmental factors， which seriously affects people’s health level and living quality. As a special research tool， the twin method can be used to estimate the relative effects of genes and environment on diabetes-related traits. Moreover， compared with regular study subjects， twins have natural controls. The extended methods combined with genetic statistics and molecular biology techniques also help to analyze risk factors of diabetes and clarify the true relationship between exposure and outcome. This paper reviewed the progress of the application of the twin method to diabetes factorial studies in the past ten years at home and abroad， and summarized the advantages and limitations of the classical twin model， co-twin control studies， and genome-wide and epigenetic epidemiological studies.