The invention relates to a power quality fuzzy evaluation method processed through an improved empowerment algorithm. According to the method, first, normalization processing is performed on all indexes of power quality; appropriate indexes are selected for evaluation, wherein the indexes which are excellent all the time are not used for evaluation, and the indexes with index values smaller than 0.1 after normalization are removed; weight calculation is performed on the selected indexes, wherein an analytic hierarchy process (AHP) is adopted to perform subjective weight calculation, an improved entropy weight method is adopted to perform objective weight calculation, and then a subjective and objective composite weight is obtained through merging; the obtained composite weight is converted into a weight angle in a radar map, the area and perimeter of the radar map are obtained through calculation in combination with all single index values, and a comprehensive index evaluation value is obtained according to the area and perimeter of the radar map; and after a comprehensive index weight value is determined, the comprehensive index weight value is substituted into a membership function, and the fuzzy evaluation method is adopted to perform grading evaluation on the obtained indexes. By means of converting single indexes into a comprehensive index through an improved radar map method, comprehensive index processing is performed on power quality, and goodness and badness of the indexes can be displayed vividly.