The internal insulation condition is a key factor to ensure its measurement accuracy and operation stability. However, during the long-term operation, the internal insulation condition would be degradation caused by electrical aging, thermal aging and system overvoltage. Aiming at this problem, a data-driven CVT internal insulation condition identification method is proposed in this paper. This method construct the measurement error feature parameters based on the electrical connection relationship between the high-voltage CVT groups, and the fuzzy analysis is utilized to match the internal insulation condition and measurement error condition of CVTs according to their correlations, then the degree and type of abnormal internal insulation could be evaluated online.