In order to provide an adequate solution in terms of robustness and automation in the process of Extract-Transform-Load (ETL) in data warehouses, in this article a multi-agent model that gathers the strengths of other approaches like wrappers and ad-hoc solutions is presented. Such a model considers the heterogeneity and availability of the data sources as well as their distributed nature. For its validation an experiment was performed using simulated and real data, which demonstrated not only its technical feasibility but also its effectiveness in terms of the percentage of processed data and the time to accomplish it.