One of the great mysteries in the Solar System is how Giant Planets formed. Two main formation scenarios coexist: disk gravitational instability and core accretion. These scenarios differ not only in the time required to form planets, but also in the final composition of the planets' interiors. In this sense, heavy element abundances are key constraints and they depend on how the ices of the planetesimal that formed the cores of these planets condensed (e.g., amorphous or crystalline).Measuring the deep oxygen abundance can help differentiating the condensation processes of the planetesimal ices. Indeed, clathration needs a larger amount of water than the amorphous ice scenario. While Galileo probably failed to measure the Jovian deep oxygen abundance, Juno should shed light on this long lasting question. Measuring Saturn's deep oxygen is a goal of the entry probe that will be proposed to ESA (Mousis et al. 2016). Regarding the Ice Giants, there is no such mission planned in the near future to measure their deep oxygen abundance and it is very challenging to probe remotely below the water cloud in these planets with microwaves. Another way to constrain the deep oxygen abundance consists in using thermochemical modeling to link upper tropospheric disequilibrium species to the deep oxygen.In this paper, we apply a thermochemical and diffusion model to the ice giant tropospheres to constrain their deep oxygen abundance from CO observations. Because the results depend on the thermal structure, on the strength of tropospheric mixing, and to a lesser extent on the deep carbon abundance, we have explored a 4D parameter space (temperature, tropospheric mixing, deep oxygen and carbon abundance) for each planet to fit their upper tropospheric composition. For instance, we have computed a series of classical thermal profiles based on dry/wet adiabats and new profiles that account for the mean molecular weight gradient at the water condensation layer (following the prescription of Leconte et al. 2016). We present the results of the 4D grids and the constraints we infer from the nominal models. These 4D grids can be used in the future once the deep temperature, tropospheric mixing and methane abundance are better known.< Réduire