Thermal fatigue of austenitic and ferritic stainless steel grades has been experimentally and numerically investigated. A special test has been developed to determine the thermal fatigue resistance of clamped V-shaped specimens. This test permits to impose thermal cycle by alternating resistance heating and air cooling. The thermal fatigue life of a specimen is expressed as the number of cycles to failure. For a given grade, the fatigue life depends on the maximal and minimal temperature of the cycle, holding time at the maximal temperature and specimen thickness. The advantage of this V-shape test is that it is a simple procedure quite representative of the thermal fatigue process occurring in an exhaust manifold. This test is well suited to perform a study of damage mechanisms and to compare stainless steel grades. Examination of the failed specimens indicated that cracks could be mainly attributed to out-of-phase (OP) thermal fatigue process especially in case of ferritic grades. For austenitic steels (AISI304 EN1.4301, AISI321 EN1.4541 or AISI308 EN1.4828) at a critical temperature or above, an in-phase (IP) thermal fatigue mechanism is coupled with oxidation and creep, which are further significantly reducing the lifetime. Therefore, the service temperature range of austenitic grades is more limited than ferritic grades. Despite their lower yield stress at high temperature, ferritic grades exhibit a very good thermal fatigue resistance at elevated peak temperatures because of their very good cyclic oxidation behavior, creep resistance and their low coefficient of thermal expansion. Consequently a dedicated titanium or niobium stabilized ferritic offer was developed for the hot part of the exhaust system (from manifold to catalytic converter) that includes 14%Cr (K11X 429/425 1.4595), 17%Cr (K41X 441 1.4509) and 19%Cr (K44X modified 444 1.4521) grades in order to cover the peak temperature range from 900°C to 1050°C.