This paper proposes a comprehensive evaluation method and a global power quality evaluation system to achieve continuous and real-time evaluation of power quality. The entropy weight method is used to construct a comparison channel for the data information between each power quality indicator, ensuring the objectivity of the weighting. State variables are also constructed to dynamically adjust the weights of each point in the calculation unit, which enhances the time sequence weighting's dynamic characteristics and improves the weight resolution in the time domain. A dual-track TOPSIS evaluation method is proposed to open up TOPSIS calculation paths for both data and grades, ensuring the consistency of evaluation standards for each period. The method also merges data into grades to obtain a comprehensive evaluation of the power quality overall state based on the evaluation grade and a cause analysis of grade changes based on the state weight. Through case analysis, this evaluation method is verified to be effective for real-time evaluation and has generality.