Abstract: The insulation aging diagnosis of capacitive voltage transformers relies heavily on single parameter threshold discrimination, which makes it difficult to effectively analyze the strong nonlinear coupling mechanism of aging characteristics under multi physical field coupling, resulting in limited early fault identification accuracy. Therefore, research on insulation aging fault diagnosis of capacitive voltage transformers based on Eclat association rule mining is carried out. By using the Eclat vertical data mining algorithm based on equivalence class transformation, strong association rules between aging features and insulation states are deeply extracted; Based on the frequent itemset iteration mechanism of Eclat association rule mining, an asymmetric confidence mapping relationship between feature items and fault types is established. Then, using the dual threshold constraints of association rule lifting degree and leverage degree, the identification and evolution path tracing of insulation aging fault types are achieved. Comparative experiments show that this method is significantly superior to traditional association analysis methods in terms of diagnostic coverage and rule redundancy suppression, verifying its robustness and engineering applicability in high-dimensional sparse data scenarios.