In Europe, rehabilitation policies for drinking water pipes are based on criteria closer to budgetary restrictions than to an authentic technico-economic consideration. In this reactive approach, the rehabilitated pipes are only pipes which have failed or have been repaired at more than a subjecive rate. At first, this paper presents different models and methods adapted to a "pro-active" approach. They can be classified into two major types according to their objectives : models assessing the structural state of the pipes based on statistical methods and models assessing and comparing the rehabilitation strategies, based on technical and economic assumptions. The first type use short or long-term data maintenance records. The second requires a description of the network (age, length, material and failure rate) and integrates some economic data. Previous studies on European cases are used to classify data requirements. For each model, we define "required data", "highly significant data" and "useful data". Previous results are given as illustration. In the second part we present case studies about available data in European services. It shows that available databases remain incomplete regarding failure analysis and modelling and rehabilitation forescasting.