Contract power conversion factor which is applied to estimate contract power of industrial customers is an important standard to caculate transformer capacity. This paper shows a reasonable contract power conversion factor, that was made by the systematic and statistical way considering actual conditions, such as investigated contact power and peak power for the last 5 years of each customer for 349 industrial customers as to AMR. In this dissertation, it is necessary to analyze the key features and general trend from the investigated data. It made an analysis of the feature parameters, such as average, standard deviation, median, maximum, minimun and thus it was carried by the linear and nonlinear regression analysis. Therefore, this paper compared characteristics for a contract power conversion factor which is a lied to calculate contract power with characteristics for a regression model for customers which maximum utilization factor of transformer is more than 60(%).