More and more human-made space objects re-enter the atmosphere, and yet the risk for human population remains often unknown because predicting their reentry trajectories is formidably complex. While falling back on Earth, the space object absorbs large amounts of thermal energy that affects its structural integrity.It undergoes strong aerodynamic forces that lead to one or several breakups. Breakup events have a critical influence on the rest of the trajectory are extremely challenging to predict and subject to uncertainties. In this work, we present an original model for robustly predicting the breakup of a reentering space object. This model is composed of a set of individual solvers that are coupled together such as each solver resolves a specific aspect of this multiphysics problem. This paper deals with two levels of uncertainties. The first level is the stochastic modelling of the breakup while the second level is the statistical characterization of the model input uncertainties. The framework provides robust estimates of the quantities of interest and quantitative sensitivity analysis. The objective is twofold: first to compute a robust estimate of the breakup distribution and secondly to identify the main uncertainties in the quantities of interest. Due to the significant computational cost, we use an efficient framework par-* Corresponding author. ticularly suited to multiple solver predictions for the uncertainty quantification analysis. Then, we illustrate the breakup model for the controlled reentry of an upper stage deorbited from a Geo Transfer Orbit (GTO), which is a classical Ariane mission.