Abstract: In this paper, in order to make the results of peak and valley time division reflect the load difference of each period objectively and be applicable in a long period of time (e.g., 1 year), a time division model is presented by combining processing of high-dimension data sampling set and K-means clustering analysis. First of all, a data sample set covering all load information within a long period of time (e.g., 1 year) is built by using the processing method of high-dimension data. Secondly, the peak and valley time division model based on the high-dimension data sample set is built by using K- means clustering analysis. Finally, the numerical simulation of proposed model is carried out by combining the load data of the whole year in certain district, and the final division result of the peak and valley time can be output on the basis of verifying the rationality of the model.