To improve the objectivity and accuracy of power quality evaluation, a grading evaluation model of power quality based on criteria importance through intercriteria correlation (CRITIC) and an improved grey-technique for order preference by similarity to an ideal solution (Grey-TOPSIS) method is proposed. First, the evaluation indices are selected based on the interaction between subsystems in a pressure-state-response (PSR) model, and a comprehensive evaluation index system is established. Because there is a certain correlation between indices, and considering the differences between indices, the CRITIC method is used to ensure the objectivity of index weighting based on two aspects of information. Then, an improved Grey-TOPSIS evaluation model is established, and the Euclidean distance is improved by introducing grey correlation degree. The new distance measure is constructed by fusing the progress of graph pasting with spatial position to make up for the defect of the original criterion. Moreover, the power quality grade is added to the evaluation matrix, and the grading evaluation of power quality is realized by judging the distance between the grade sequence and the positive ideal solution. The case study indicates that the improved model makes full use of data information, reduces the impact of subjective factors, and can obtain more accurate and comprehensive power quality evaluation results. This work is supported by the National Key Research and Development Program of China (No. 2018YFE0122200).