Background. Nonalcoholic fatty liver diseases are often associated with obesity, insulin resistance, and excessive visceral fat accumulation. The aims of this study were (1) to evaluate the relationship between the severity of fatty liver and visceral fat accumulation in nonalcoholic fatty liver diseases, and (2) to investigate the relationships of fatty liver with biochemical data and insulin resistance. Methods. One hundred twenty-nine subjects (63 women) with fatty liver diagnosed by ultrasonography were enrolled. Subjects positive for hepatitis B virus, hepatitis C virus, or autoimmune antibodies and those whose alcohol intake was over 20g/day were excluded. The visceral fat area at the umbilical level and the liver-spleen ratio were evaluated by computed tomography. Results. The severity of fatty liver evaluated by ultrasonography showed a significant positive relationship with the visceral fat area and waist circumstance (fatty liver severity: mild, 92.0 ±30.9 cm2; moderate, 122.1 ±32.6 cm2; severe, 161.0 ±48.4 cm2; P < 0.0001). The visceral fat area and liver-spleen ratio were negatively correlated (r = -0.605, P < 0.0001). The severity of fatty liver showed strong positive relationships with serum aspartate aminotransferase, alanine aminotransferase, fasting plasma glucose, fasting plasma insulin, and insulin resistance. The severity of fatty liver was positively related to the visceral fat area in 49 nonobese subjects (body mass index < 25). Conclusions. The severity of fatty liver was positively correlated with visceral fat accumulation and insulin resistance in both obese and nonobese subjects, suggesting that hepatic fat infiltration in nonalcoholic fatty liver disease may be influenced by visceral fat accumulation regardless of body mass index. Background. Nonalcoholic fatty liver diseases are often associated with obesity, insulin resistance, and excessive visceral fat accumulation. The aims of this study were (1) to evaluate the relationship between the severity of fatty liver and visceral fat accumulation in nonalcoholic fatty liver diseases. One hundred twenty-nine subjects (63 women) with fatty liver diagnosed by ultrasonography were enrolled. Subjects positive for hepatitis B virus, hepatitis C virus, or autoimmune antibodies and those whose alcohol intake was over 20g / day were excluded. The visceral fat area at the umbilical level and the liver-spleen ratio were evaluated by computed tomography. Results. The severity of fatty liver evaluated by ultrasonography showed a significant positive relationship with the visceral fat area and waist circumstance (fatty liver severity: mild, 92.0 ± 30.9 cm2; moderate, 122 The visceral fat area and liver-spleen ratio were negatively correlated (r = -0.605, P <0.0001). The severity of fatty liver showed strong positive relationships The severity of fatty liver was positively related to the visceral fat area in 49 nonobese subjects (body mass index <25). Conc severity. The severity of fatty liver was positively correlated with visceral fat accumulation and insulin resistance in both obese and nonobese subjects, suggesting that hepatic fat infiltration in nonalcoholic fatty liver disease may be by visceral fat accumulation regardless of body mass index.