In this paper, we propose a forecasting system for a multi-period assignment model, where Internet prospects' flows evolve continuously over time. In particular, we show how we build a forecasting model of Internet prospects' flows while considering some characteristics of a multi-period assignment model in order to optimize and improve the marketplace's system of Place des leads (PdL). A numerical investigation is conducted based on real data of PdL application to show the benefits of introducing forecasting information to the multi-period optimization model