In this paper, power quality regionalization control method with improved piecewise linear representation (IPLR) and improved gray relational analysis (IGRA) was proposed to improve the power quality and operation efficiency with big power quality data. According to the variations of adjacent and interval sampling points of time series, the characteristic points of power quality time series are extracted. The IPLR is applied to reduce the dimension of sequence. With the processed sequence, the coupling relationship of power quality between nodes can be quantified into correlation degree by IGRA method. Based on the correlation degree, a regionalization strategy was defined to figure out the regionalization results and master nodes by hierarchical clustering method. With the proposed method, the power quality control can be realized by regionalization control and the algorithm efficiency can be improved. The standard IEEE 14 bus system simulation was constructed to verify the effectiveness of the proposed method.