Representing and processing digital data related to underground networks, particularly sewerage networks, is increasingly becoming a priorityfor the managers of these networks. Indeed better representation would allow them, among others, to improve knowledge and to take the best decisions regarding these generally poorly identified infrastructures. The heterogeneity of data and the multiplicity of data models representing sewerage networks, often specific to each operator, as well as the imperfections associated with both the available data and those collected from different sources, generate complexity in terms of on-the-field interventions efficiency. They also highlight the need for aggregation (unification),control and analysis. The main objective of our work is to merge multisource data to obtain more precise and complete digital maps of sewerage networks. In this paper, we propose a generic data modelling for dataf usion purposes taking into consideration the uncertainty aspects related to the collected data by allowing a confidence value for each data source and for each single data provided by a source.