BACKGROUND: Alzheimer's disease (AD) is a progressive and chronic neurodegenerative disorder of the central nervous system, characterized by behavioral and dysexecutive deficits. Its pathogenesis is closely associated with the intestinal flora. This study aimed to explore the enterotypes in AD by identifying key bacteria through machine learning and species co-occurrence network analysis. METHODS: The collection of fecal samples from AD patients was followed by 16 S rRNA analysis using QIIME2. Enterotype clustering was conducted at the genus level, and deep neural network (DNN) classification models were developed for AD and healthy controls within each enterotype. RESULTS: Analysis of three 16 S rRNA gut microbiome datasets identified three distinct enterotypes: Escherichia_Shigella (ET-E), Faecalibacterium (ET-F), and Bacteroides (ET-B). The ET-E is mainly characterized by the absence of Akkermansia in AD group. The Akkermansia was significantly positively correlated with Eubacterium_coprostanoligenes_group and negatively correlated with biosynthesis and amino acid metabolism. The ET-F highly expressed Agathobacter, un_f__Lachnospiraceae, Lachnoclostridium, and low expressed Dorea in AD group. Among them, Agathobacter was significantly positively correlated with un_f__Lachnospiraceae, and un_f__Lachnospiraceae was significantly positively correlated with Lachnoclostridium. The Dorea was significantly negatively correlated with Lachnoclostridium. The AD from ET-B group had high expression of two beneficial bacteria, Butyricicoccus and Parabacteroides. The findings suggest that the ET-E enterotype may predispose individuals to AD, with Akkermansia identified as a potential risk factor. Conversely, the ET-B enterotype appears to be associated with milder symptoms, with Butyricicoccus and Parabacteroides potentially serving as protective factors. Therefore, a comprehensive understanding of the species characteristics and interactions within different enterotypes is essential for modulating the gut-brain axis and mitigating AD symptoms.