Traditional selecting optimization structures at large table in data warehouse can only either shorten the time of query processing, or reduce the storage and maintenance overhead. It can't achieve a combined effect at the same time. Based on the traditional optimization techniques, this paper proposes a new approach, namely, HP→BIs, by exploiting their similarities of horizontal partitioning and bitmap indexes. Experimental results show that with this approach, we can significantly shorten the time of query processing, reduce the storage and maintenance cost.