Alzheimer＇s　disease　(AD)　is　a　neurodegenerative　disease　that　cannot　be　reversed　once　it　occurs,　and　there　is　currently　no　cure.　However,　early　detection　of　AD　can　lead　to　earlier　control　of　the　patient＇s　condition,　thus　maximizing　their　quality　of　life.　This　study　constructed　a　brain　network　graph　based　on　the　disease　characteristics　of　AD　and　scalp　electroencephalogram　(EEG)　data,　and　extracted　topological　features.　These　features　were　then　input　into　machine　learning　and　deep　learning　models　for　training,　achieving　a　maximum　accuracy　rate　of　97.27%.　This　suggests　that　brain　network　structures　can　be　used　to　identify　AD,　which　may　help　doctors　diagnose　AD　earlier.　©　2023　IEEE.