Abstract: Accurate evaluation for corona resistance life time of enamelled wires is always been a challenging issue in the winding wires industry due to the large dispersion of test data. This large dispersion makes it difficult for us to determine whether there is significant difference between each laboratory. And the determination of dispersion patterns of within-group data is prior to the analysis on the significant difference of between-group data. This paper uses three typical life distribution functions such as normal distribution, lognormal distribution and Weibull distribution to analyze the data of corona resistance life of a kind of enamelled wires by linear regression method. Results showed that the correlation coefficients of three types of functions were ranked as three-parameter Weibull distribution, log-normal distribution and normal distribution. Therefore, the three-parameter Weibull distribution could give a better description of the distribution characteristics of the data for corona resistance life time of enamelled wires, while the normal distribution is not suitable. Lognormal distribution model gives a shorter characteristic value of corona-resistance life time at low failure probability and a higher one at high failure probability, which shows that Lognormal distribution model is not suitable for the calculation of corona resistance life time at high or low failure probability. However, three-parameter Weibull distribution model shows more advantages in accuracy and safety.