The aim of this study is to explore the effect of genetic variation on diabetic retinopathy (DR) risk in a Taiwanese population. The logistic regression model was used to evaluate the relationship between DR status and risk factors, including the conventional parameters and genetic risk score (GRS). Candidate single nucleotide polymorphisms (SNPs) in GRS were selected based on previous reports with a combined P < 10−4 (genome-wide association) and P < 0.05 (meta-analysis). In total, 58 SNPs in 44 susceptibility loci were selected, and four were used to calculate GRS. After adjustment for age, systolic blood pressure, diabetes duration, and HbA1c, the DR risk was 4.95 times higher for patients in the top GRS third tile than for those in the bottom third tile (95% CI = 2.99–8.18; P < 0.001). The addition of genetic information improved DR prediction, increasing the area under the curve (AUC) from 0.72 to 0.77 (P = 0.0024) and improving the sensitivity of the model such that 40 more subjects were reclassified into DR status. The developed multivariate logistic regression model combining conventional risk factors and the multilocus GRS can predict DR, thus enabling timely treatment to reduce blindness in T2D patients.