In China, as educational reforms progress, the characteristics of teachers' work have undergone significant changes, resulting in extremely high levels of stress that can trigger anxiety and depression. Anxiety and depression often co-occur, with two mainstream theories explaining this co-existence: the tripartite model and the diathesis-stress model. However, systematic research focusing on this population is relatively scarce, and the applicability of these models has not been thoroughly tested. This study aims to use network analysis methods to examine the interactions between symptoms and analyze the co-existence of anxiety and depression, thereby expanding the research on teachers. Data were provided by the Science Database of People Mental Health, which includes 1670 teachers with a mean age of 30.01. The Self-Rating Anxiety Scale and Self-Rating Depression Scale were used to estimate the network structures of anxiety and depression, respectively. Shared symptoms between depression and anxiety were identified using network analysis and clique percolation methods. Bayesian Networks was used to estimate causal relationships between symptoms. Data were analyzed using R packages. Network structure was constructed with the qgraph package, node centrality and bridge symptoms were evaluated using the networktools package, and network stability was measured via the bootnet package. The Clique Percolation method was implemented with the CliqurPercolation package, and Bayesian network modeling was performed via the Bnlearn package. Dizziness and Easy Fatigability & Weakness were central symptoms in the network. Bridging strength results showed that, the important bridging symptoms included Tachycardia, Depressed Affect, Fatigue, Crying Spell, Easy Fatigability & Weakness, Nightmares, Face Flushing, and Sweating were the strong bridging symptoms. Additionally, Sleep Disturbance played a key mediating role. Depressed Affect and Dissatisfaction were activation symptoms for anxiety-depression co-existence. Using network analysis, this study elucidated core, bridging, and shared symptoms, as well as potential causal pathways between anxiety and depression. Specifically, somatic symptoms are crucial in maintaining and developing the anxiety-depression network among teachers. Sleep disturbance serves as the sole gateway for mild symptoms to develop into other communities. The Bayesian network identified two key activating symptoms within the teacher anxiety-depression network, validating the applicability of the tripartite model among teachers.