Closing the nuclear fuel cycle is essential for the development of low-carbon energy systems. In France, plutonium resulting from the reprocessing of irradiated UOX fuels is recycled through the fabrication of MOX (U,Pu)O2 fuel and used in some pressurized water reactors (PWRs). MOX fuel is also envisaged for future Generation IV sodium-cooled fast reactors (SFRs), including some concepts of small modular reactors (SMRs). In these reactors, the strong radial thermal gradient modifies the fuel microstructure and leads to a radial redistribution of oxygen, and therefore of the O/M ratio (M = U + Pu). The safe deployment of this fuel cycle therefore relies on obtaining new data, currently inaccessible, on the properties that govern heat transfer in MOX, notably thermal conductivity (lambda, W.M-1.K-1) and heat capacity (Cp, J.kg-1.K 1). Indeed, the available data remain limited for sub-stoichiometric MOX and for irradiated MOX. Fuel performance codes (FPCs) currently use correlations established from data obtained on UO2 or on MOX within restricted composition and temperature ranges. These extrapolations are particularly limiting for describing the hot zones of irradiated MOX, where the O/M ratio can reach values inaccessible by direct fabrication of a single-phase MOX. This PhD work therefore aims to study the impact of the O/M ratio and of the microstructure on the thermophysical properties of MOX fuel, currently inaccessible, between ~500 and ~3000 K. The measurements will be carried out by laser flash on MOX fabricated according to the RNR and COCA processes, in order to compare the impact of the different initial microstructures resulting from these two fabrication routes. In order to study and reproduce the most sub-stoichiometric states observed in reactor, an innovative strategy of in-situ adjustment of the O/M ratio will be developed in the CLASH device. The PhD will take place at CEA Marcoule, where the characterizations of the materials before and after thermophysical measurements will be carried out which will, for their part, be performed at JRC Karlsruhe (Germany) during a 12-month mobility. In parallel, thermodynamic calculations (CALPHAD) will be carried out in order to interpret the phase equilibria at high temperature, and thus relate the measured properties to the structural and physico-chemical evolutions of MOX. These new data will feed the FPCs in order to more precisely evaluate the temperature and the margin to MOX melting.