The data warehouses (DW) are proposed to collect and store heterogeneous and bulky data. They represent a collection of thematic, integrated, non-volatile and histories data. They are fed from different data sources through transactional queries and offer analytical data through decisional queries. Generally, the decisional queries execution cost on large tables is very high. Reducing this cost becomes essential to enable decision-makers to interact in a reasonable time. In this context, DW administrators use different optimization techniques such as fragmentation, indexing, materialized views, and parallelism. On the other hand, the volume of data residing in the DW is constantly evolving. This can increase the complexity of frequent queries, which can degrade the performance of DW. The administrator always has to manually design a new fragmentation scheme from the new load of frequent queries. Having an automatic fragmentation tool of DW becomes important. The approach proposed in this paper aims at an incremental horizontal fragmentation technique of the DW through a web service. This technique is based on the updating of the queries load by adding the new frequent queries and eliminating the queries which do not remain frequent. The goal is to automate the implementation of the incremental fragmentation in order to optimize the new queries load. An experimental study on a real DW is carried out and comparative tests show the satisfaction of our approach.