Abstract: Aiming at that the existing density peak fast search algorithm does not consider the temporality and cannot process the dynamic time sequence data, the time series weighting factor is added on the basis of the density peak fast search algorithm to improve the topological relationship among data points and proposed time series weighted density peak clustering algorithm, which makes the density peak fast search algorithm have the ability to deal with dynamic time series data. Using the user load classification model based on the time series weighted density peak clustering algorithm to process the user power load data set published by OpenEI, the clustering effect is more accurate than the user load classification model based on the density peak clustering algorithm and the statistical evaluation index have improved.