Objective: Several single-nucleotide polymorphisms have been identified to be disadvantageous or protective in regard to disease severity in patients with non-alcoholic fatty liver disease (NAFLD). However, it is unclear, whether including genetic risk factor(s) either alone or combined into risk stratification algorithms for NAFLD actually provides incremental benefit over clinical risk factors. Design: Patients with biopsy-proven NAFLD were genotyped for the PNPLA3-rs738409(minor allele:G), TM6SF2-rs58542926(minor allele:T) and HSD17B13- rs72613567 (minor allele:TA) variants. The NAFLD activity score (NAS) and fibrosis stage (F0–F4) were used to grade and stage all liver biopsy samples. Patients from seven centers throughout Central Europe were considered for the study. Results: 703 patients were included: NAS ≥ 5:173(24.6%); Fibrosis: F3–4:81(11.5%). PNPLA3 G/G genotype was associated with a NAS ≥ 5(aOR 2.23, p = 0.007) and advanced fibrosis (aOR-3.48, p < 0.001).TM6SF2 T/- was associated with advanced fibrosis (aOR 1.99, p = 0.023). HSD17B13 TA/- was associated with a lower probability of NAS ≥ 5(TA/T: aOR 0.65, p = 0.041, TA/TA: aOR 0.40, p = 0.033). Regarding the predictive capability for NAS ≥ 5, well-known risk factors (age, sex, BMI, diabetes, and ALT; baseline model) had an AUC of 0.758, Addition of PNPLA3(AUC 0.766), HSB17B13(AUC 0.766), and their combination(AUC 0.775), but not of TM6SF2(AUC 0.762), resulted in a higher diagnostic accuracy of the model. Addition of genetic markers for the prediction of advanced fibrosis (baseline model: age, sex, BMI, diabetes: AUC 0.777) resulted in a higher AUC if PNPLA3(AUC 0.789), and TM6SF2(AUC 0.786) but not if HSD17B13(0.777) were added. Conclusion: In biopsy-proven NAFLD, PNPLA3 G/-, TM6SF2 T/- and HSD17B13 TA/- carriage are associated with severity of NAFLD. Incorporating these genetic risk factors into risk stratification models might improve their predictive accuracy for severity of NAFLD and/or advanced fibrosis on liver biopsy.