To better control the time, cost and resources assigned to software projects, organizations need a proper estimate of their size even before the projects actually start. Accordingly, several approaches were proposed to estimate the size of a software project, as the well-known Function Point Analysis (FPA) , which is largely used in traditional software development projects. However, we observed in our company that it is not fit for data mart software measurement. Data Mart (DM) systems have particularities in their development that are different from the traditional software systems (e.g. a DM uses other software systems as data sources and does not create new information). It is important, therefore, to have a measurement approach that considers those particularities while measuring the DM size. In this paper, we present an adaptation of the FPA approach for DM size measurement and discuss results on 10 data marts project developed in the industry.