The invention, which relates to the power grid field, provides a data-division-based hierarchical clustering method for a power load curve. On the basis of a clustering effectiveness evaluation function DBI, clustering is carried out on power load curve data by using a data-division-based hierarchical clustering algorithm; and clustering time is reduced by using a threshold setting way. Because the data are divided into a plurality sub sets and then clustering is carried out separately, the clustering time of the power load curve is greatly shortened. The multiple experiments show that when the DBI index is used for evaluating the clustering quality, the data-division-based hierarchical clustering algorithm enables the clustering quality to be improved by about 3% by being compared with traditional hierarchical clustering algorithm.