Abstract: As major engineering structures evolve towards larger scales, increased complexity, and operation in extreme environments, the evaluation of material service safety faces challenges such as multi-factor coupling, cross-scale damage evolution, and long-term performance degradation. Traditional methods relying solely on data-driven or physical mechanism models struggle to meet the high precision and reliability requirements of evaluations. Consequently, collaborative evaluation methods based on the integration of data and mechanisms have become a current research hotspot.This paper systematically reviews the research progress of data-mechanism fusion modeling methods in the service safety evaluation of major engineering materials. It summarizes various approaches to material service safety evaluation, including statistical methods, artificial intelligence modeling methods, and data-mechanism fusion methods. The study reveals that the integration of data-driven and mechanism models holds broad application prospects in life prediction under complex environments, enhancing the accuracy and credibility of predictions and analyses. Looking ahead, further enrichment of material gene databases, research on high-precision machine learning algorithms, optimization of model parameters, improvement of computational efficiency, and the integration of physical models and related theories will contribute to advancing the application of deeply integrated data and mechanism methods in the service safety evaluation of major engineering materials.