Abstract: Conventional harmonic responsibility determination methods have shortages such as the inability to achieve long-time scale dynamic responsibility determination, the need for special measurement, the inability to utilize the data of existing monitoring systems, and the inability to determine the harmonic responsibilities of users on the same feeder. Thus, this paper proposes a determination method of user harmonic responsibility based on correlation analysis of monitoring data. Firstly, the monitoring data are collected, including the harmonic voltage data from the power quality monitoring system and the average active power data of each user from the power information collection system. By the canonical correlation coefficients based on these data, the harmonic responsibility at the point of common coupling（PCC） is divided into the background harmonic responsibility and the harmonic responsibility of concerned users. Secondly, the dynamic distortion correlation coefficients of harmonic voltage and active power of each user are calculated, which are used to reflect the relationship between user power consumption behavior and harmonic voltage distortion at PCC. Finally, the long-time scale index for harmonic responsibility determination is developed with consideration of the relationship and user power consumption capacity. The validity and practicability of the proposed method are verified by the data in the actual monitoring system, and the long-time scale dynamic harmonic responsibility determination of each user on the feeder can be realized based on the existing monitoring data.