The aim of this study was to evaluate the effects of ursodeoxycholic acid on patients with nonalcoholic steatohepatitis using meta-analysis. PubMed, EMBASE, Web of Science, Cochrane Library, Chinese Biomedical Databases, and article references were searched. We included randomized controlled trials using liver biopsy as a reference standard. We identified three eligible studies. Among histological responses, only lobular inflammation improved in the high-dose ursodeoxycholic acid subgroup compared with the control group [mean deviation (MD): -0.23 (-0.40, -0.06), P=0.008]. However, fibrosis may tend to increase [MD: 0.08 (-0.04, 0.20), P=0.17]. Among biochemical responses, γ-glutamyl transpeptidase reduction was significantly greater in the ursodeoxycholic acid group than in the placebo group, and the reduction tendency was only shown in the high-dose subgroup [MD: -35.58 (-52.60, -18.56), P<0.0001]. Serum total bilirubin increased in the high-dose ursodeoxycholic acid subgroup compared with the control group [MD: 0.43 (0.14, 0.72), P=0.004]. Ursodeoxycholic acid-treated patients did not differ significantly from control patients with regard to alanine transaminase, aspartate aminotransferase, and alkaline phosphatase activities. Adverse events were nonspecific and considered of no major clinical relevance. Ursodeoxycholic acid in monotherapy has no substantial positive effect on nonalcoholic steatohepatitis.