Nonalcoholic fatty liver disease (NAFLD) is a metabolic liver disease with widely variable phenotypes extending from simple steatosis, through nonalcoholic steatohepatitis (NASH) to cirrhosis and hepatocellular carcinoma. Inevitably, this reflects the interplay of well-recognized environmental factors and disease associations such as obesity and insulin resistance with host genetic factors, which are polygenic or complex in nature. Most of the observed phenotypic variability will probably be explained by variations in single nucleotide polymorphism frequency, although knowledge of the effect of most polymorphisms on biologic function is currently limited. Several observational studies of kindred with NASH suggest a genetic contribution. Most data characterizing genetic variation in different NAFLD phenotypes is derived from case-control association studies involving putative candidate genes. These candidate genes have been selected largely based upon the "two-hit hypothesis" of the pathogenesis of NAFLD, although other hypothesis-independent approaches can also be informative in gene selection. Thus far, candidate gene association studies have had significant limitations such as small cohort sizes and poor reproducibility. Rapid technologic developments are increasing the capability of detecting genetic variation. Identification of the genetic contribution to NAFLD will inform theories of disease pathogenesis and progression and ultimately improve management.