Requirements of efficiency and safety have steadily increased across power grid in recent years. In terms of 35 kV and above capacitive voltage transformer (CVT), the state of the equipment is evaluated by testing the insulation resistance, capacitance and dielectric loss factor of each part of the equipment at present. However, there are periodic outage tests, which fail to meet the requirements for efficient operation and maintenance of equipment. Online monitoring technology can realize the status assessment in the operation of power equipment. The current CVT secondary voltage online monitoring technology suffers from low accuracy in alarm detection during application, leading to frequent misjudgments and missed detections of equipment defects. In this study, CVT operation data, switching signal and network topology information are used to construct CVT defect identification and analysis model and state criterion in complex power grid. The data utilized originates from the energy management system (EMS), and the collected information encompasses measuring point information, measuring point values, time points, real-time load data, and switch state data. A CVT consists of a capacitor unit and an electromagnetic unit. The capacitor unit is further divided into a high voltage capacitor, C1, and a partial voltage capacitor, C2. Both C1 and C2 are composed of multiple capacitor components connected in series. Therefore the equivalent circuit model of the CVT has been constructed, and the relationship between the secondary voltage and the high-voltage capacitance as well as the partial voltage capacitance has been analyzed. The results indicate that the secondary output voltage of CVT increases when there is a breakdown in the capacitor component of the high voltage capacitor C1. Conversely, the secondary output voltage of CVT decreases when there is a breakdown in the capacitor component of the partial voltage capacitor C2. The breakdown of capacitive elements is a gradual process, not a sudden event. Consequently, on-line monitoring of equipment capacity can be accomplished through secondary voltage monitoring. The criteria for assessing the condition of CVT equipment comprise two categories: general criteria and specific criteria. Among them, the general criteria apply to all CVT in the power grid, including the absolute value of the secondary voltage change rate, the secondary voltage unbalance rate of three-phase CVT, and the cumulative value of alarm times. The special criterion involves conducting a differentiated state assessment based on the various positions of CVT within the power grid. For the line CVT, the auxiliary criterion includes the line switch state and CVT voltage difference between the first and last ends of the line. For the CVT at the bus junction, the auxiliary criterion increases the secondary voltage unbalance rate of all CVT devices on the same bus and the voltage difference from the same bus. After 7 months of pilot application by X Power Supply Company, the number of alarm defect misjudgments in 35kV and above CVT online monitoring has decreased from 25 to 7, and the CVT alarm accuracy rate has increased from 63 to 90%. The research results can realize the dynamic perception of CVT state in operation and accurate warning of defects in complex power grid, and guide on-site operation and maintenance.