The management of big databases of three- dimensional models (used in CAD applications, visualization, games, etc.) is a very important domain. The ability to characterize and easily retrieve models are a key issues for the designers and the final users. In this frame, two main approaches exist: search by example of a three-dimensional model, and search by a 2D view. In this paper we focus on the characterization of a 3D model by a set of views (called characteristic views), and on the indexing process of these models with a Bayesian probabilistic approach using the characteristic views. We illustrate our results using a collection of three- dimensional models supplied by Renault Group.