Objective: To explore the applied characteristics of back-propagation artificial neural network (BPANN) on studying the genetic variants in PPAR-γ and RXR-α gene and type 2 diabetes risk in a Chinese Han population. Methods: With BPANN as fitting model based upon data gathered from type 2 diabetes patients(n=591) and normal controls (n=724), the mean impact value (MIV) for each input variables and sequencing the factors according to their absolute MIVs were calculated. The results from BPANN were compared with multiple logistic regression analysis,and multifactor dimensionality reduction (MDR) method was used to consider the joint effects of PPAR-γ and RXR-α gene. Results:By BPANN analysis, the risk factors of diabetes mellitus were serum adiponectin level, hypertension, waist, rs4240711, rs3132291, rs3856806, diabetes mellitus family history,no alcohol drinking hyperlipoproteinmia, smoking, age, body mass index, rs1045570, gender, rs2920502, rs6537944, rs4842194, rs17827276 and rs1801282. However, only 8 factors were statistically significant in multiple logistic regression analysis, arrayed according to the important valne: hypertension, diabetes mellitus family history, waist, no alcohol drinkiy, smoking, rs4240711, rs4842194 and serum adiponectin level. Model X1 X2 X3(rs3856806, rs3132291, rs4240711) was the best model (CV Consistency=10/10, P=0.0107) with MDR method. Conclusion: These results suggested that the interactions of PPAR-γ and RXR-α gene might have important role in the susceptibility of T2DM. Neural network could be used to analyze the risk factors of diseases and more complicated relationships (main effects and interactions) between inputs and outputs, better than using the traditional methods.