The invention discloses an electricity price setting method and system. The electricity price setting method includes the steps: performing gaussian smoothing processing on load data matrixes of a plurality of users by means of gaussian kernel function, and obtaining a total load data matrix; randomly extracting k smooth load matrixes from the total load data matrix as the initial clustering centers, and obtaining a clustering center set; by means of a K-means algorithm, calculating the distance of the smooth load matrix of each user and each clustering center, and utilizing the improved similarity identification function to classify each load data matrix to each initial clustering center; and utilizing a center point calculation formula to recalculate the clustering centers to perform multiple times of iteration clustering, and obtaining k user groups which are divided according to the initial clustering centers; by means of a time-of-use electricity price decision model, on the basis of a weighting method, solving the target function, and obtaining the respective time-of-use electricity price of k user groups, wherein the improved similarity identification function is shown in the description. The electricity price setting method and system can improve the clustering identification effect, enable the user groups to be relatively more accurate, and can more accurately set price.