Abstract： Background::Anorexia nervosa (AN) is a psychological disorder, which is characterized by the misunderstanding of body image, food restriction, and low body weight. An increasing number of studies have reported that the pathophysiological mechanism of AN might be associated with the dysbiosis of gut microbiota. The purpose of our study was to explore the features of gut microbiota in patients with AN, hoping to provide valuable information on its pathogenesis and treatment.Methods::In this cross-sectional study, from August 2020 to June 2021, patients with AN who were admitted into Peking University Third Hospital and Peking University Sixth Hospital ( n = 30) were recruited as the AN group, and healthy controls (HC) were recruited from a middle school and a university in Beijing ( n = 30). Demographic data, Hamilton Depression Scale (HAMD) scores of the two groups, and length of stay of the AN group were recorded. Microbial diversity analysis of gut microbiota in stool samples from the two groups was analyzed by 16S ribosomal RNA (rRNA) gene sequencing. Results::The weight (AN vs. HC, [39.31 ± 7.90] kg vs. [56.47 ± 8.88] kg, P < 0.001) and body mass index (BMI, AN vs. HC, [14.92 ± 2.54] kg/m 2vs. [20.89 ± 2.14] kg/m 2, P < 0.001) of patients with AN were statistically significantly lower than those of HC, and HAMD scores in AN group were statistically significantly higher than those of HC. For alpha diversity, there were no statistically significant differences between the two groups; for beta diversity, the two groups differed obviously regarding community composition. Compared to HC, the proportion of Lachnospiraceae in patients with AN was statistically significantly higher (AN vs. HC, 40.50% vs. 31.21%, Z = –1.981, P = 0.048), while that of Ruminococcaceae was lower (AN vs. HC, 12.17% vs. 19.15%, Z = –2.728, P = 0.007); the proportion of Faecalibacterium (AN vs. HC, 3.97% vs. 9.40%, Z = –3.638, P < 0.001) and Subdoligranulum (AN vs. HC, 4.60% vs. 7.02%, Z = –2.369, P = 0.018) were statistically significantly lower, while that of Eubacterium_hallii_group was significantly higher (AN vs. HC, 7.63% vs. 3.43%, Z = –2.115, P = 0.035). Linear discriminant effect (LEfSe) analysis (LDA score >3.5) showed that o_Lachnospirales, f_Lachnospiraceae, and g_Eubacterium_hallii_group (o, f and g represents order, family and genus respectively) were enriched in patients with AN. Microbial function of nutrient transport and metabolism in AN group were more abundant ( P > 0.05). In AN group, weight and BMI were significantly negatively correlated with the abundance of Bacteroidota and Bacteroides, while positively correlated with Subdoligranulum. BMI was significantly positively correlated with Firmicutes; HAMD scores were significantly negatively correlated with Faecalibacterium.Conclusions::The composition of gut microbiota in patients with AN was different from that of healthy people. Clinical indicators have correlations with the abundance of gut microbiota in patients with AN.