The increase in engine power has resulted in a significant rise in both thermal and mechanical loads on cylinder heads, leading to the emergence of thermomechanical fatigue (TMF) as one of the significant issues. This paper proposed a parameter identification method for the Neu/Sehitoglu model, utilizing multi-objective optimization (MOO), and applied it to predict the TMF life of compacted graphite cast iron (CGI) material used in cylinder heads. The study assessed the TMF behavior of RuT450 material through various experiments, including tensile, low cycle fatigue (LCF), creep, TMF, and oxidation experiments. Employing the MOO, the unknown model parameters are estimated and optimized to predict the TMF life of RuT450 materials. Results demonstrated that over 95% of the predicted TMF life falls within twice the error band of experimental values. Furthermore, plastic strain amplitude, temperature, and strain rate significantly influenced the damage mechanism and life prediction. The TMF life under out-of-phase (OP) loading conditions was found to be smaller than that of in-phase (IP) loading. Notably, under out-of-phase thermomechanical fatigue (OP TMF), oxidation damage and fatigue damage together contributed to over 95% of the total damage in RuT450 materials. Moreover, variations in oxidation damage parameters had a considerable impact on the predicted life. This study might offer an effective method for determining Neu/Sehitoglu model parameters and establish material boundaries for TMF life assessment of cylinder heads.