The use of Carbon/Carbon (C/C) ceramic matrix composites in aeronautics braking systems is justified by their high-temperature resistance (up to 1000°C) and brake friction properties. The C/C considered in this study is a 2.5D needle-punched and infiltrated by a CVI process. The preform is made of carbon yarn sheets of ex-PAN, stacked in [0/±60] and needle-punched in the transverse direction. During the landing phase, the carbon heat sink that makes up the brake reaches nominal temperatures ranging from 400°C to 800°C. Within this temperature range, carbon reacts significantly with the oxygen present in the air. Furthermore, during taxi-out and the disembarkation phase, even if the brakes are less mechanically stressed, they remain hot before reaching an acceptable temperature for take-off. So, this study aims developing a model for the specific lifetime of C/C, including the effects of carbon oxidation on thermo-mechanical properties. A literature review has provided extensive information on the behaviour of C/C composites at room temperature and from 400°C to 1000°C during carbon oxidation. The conventional measure of oxidation, expressed as the loss of mass relative to the initial mass (Δ), is commonly used not only to characterize the material's state but also to assess its residual capabilities. However, for C/C composites, previous works [1] suggest that the influence of oxidation on mechanical behaviour differs depending on the temperature at which it occurs. It indicates that the standard description of oxidation is inadequate to explain its effects on mechanical properties. The modelling exercise is carried out in two steps. The first is to understand the various oxidation phenomena and to find a formalism that can distinguish between them. Indeed, the fact that the oxidation temperature influences mechanical effects at iso-loss of mass suggests that multiple oxidation mechanisms are present depending on the temperature, and that their effects on mechanical properties are different. The aim is therefore to design a descriptor capable of capturing the state of the material through new internal state variables. The second consists of establishing a thermodynamically admissible behaviour law based on the usual framework of continuum damage models developed at ONERA [2, 3]. These models should be able to describe the effects of oxidation on the material behaviour which are non-negligible [4], even for relatively small mass losses (Δ < 5%), and at least as important as those resulting from mechanical loads, both. Then, this behaviour law must be implementable in an industrial computational code to carry out finite element analyses on structures, both in static and fatigue load cases.