Non-alcoholic fatty liver disease (NAFLD) is a common chronic liver condition. It is manifested by hepatic steatosis (HS) that can progress to non-alcoholic steatohepatitis (NASH), and even liver failure. Interestingly, it is marked by racial/ethnic disparities, with a high prevalence in Hispanics. We aimed to identify the risk factors for these chronic conditions in the US. To this end, we analyzed data from NHANES III (1988-1994) using multiple or multinomial logistic regression considering the design and sample weight. HS was identified by ultrasound. NAFLD was defined as HS in the absence of viral hepatitis or excessive use of alcohol or hepatotoxic drugs. The NAFLD population was further divided into those with NASH (defined by the HAIR score), or with simple NAFLD. The prevalence of HS was 19.8%, 16.6%, and 27.9%; of NAFLD was 17.8%, 14.7%, and 25.5%; and of NASH was 3.2%, 2.5%, and 5.1% in non-Hispanic Whites, non-Hispanic Blacks and Hispanics, respectively. Race/ethnicity was a significant predictor of HS, NAFLD and NASH, with Hispanics having the highest odds for all conditions, and non-Hispanic Blacks having the lowest odds relative to Whites (p<0.05). Other significant risk factors for all three conditions were older age, higher BMI, abnormal levels of C-peptide, and elevated serum glucose and triglycerides (p<0.05). HOMA insulin resistance was associated with HS and NAFLD (p<0.05). While smoking status was not associated with HS (p>0.05), current smokers had lower odds of NAFLD & NASH than non-smokers (p<0.05). Elevation of the liver enzyme aspartate aminotransferase was a significant risk factor of HS, while elevation of the liver enzyme alanine transaminase was a significant risk factor of NAFLD. Elevation in the levels of both liver enzymes was predictive of NASH (p<0.05). Although we included physical activity relative to national recommendation variable and the Healthy Eating Index (a measure of diet quality) in our analyses, neither of these factors was a predictor of any of the liver conditions (p>0.05). Our results showed an independent association between race/ethnicity and HS, NAFLD, and NASH, whereby Hispanics had the highest odds for every condition relative to non-Hispanic Whites. Providers should consider the race/ethnicity of their patients when evaluating the risk for NAFLD and NASH, and also be aware of the other risk factors, such as BMI and levels of C-peptide, glucose, and triglycerides.