Anticipatory failure determination (AFD) is a failure prediction method used to identify deficiencies and hidden risks in product design during design or manufacturing stages, serving as a powerful tool to enhance product reliability. Current AFD mainly define scenarios of successful realization of product functions at a macro level and give insufficient attention to the underlying mechanisms that drive functional realization, which affects the rationality and accuracy of failure prediction. Furthermore, the existing studies for function reversal and standard-solution reversal predominantly adopt direct semantic reversal to generate failure scenarios, which leads to relatively simplistic reversal methods. To address these issues, firstly, a mechanism-based failure prediction approach was proposed. In this approach, the mechanism model of the product was extracted, the mathematical or physical expressions of the mechanism model were used to query potential failures, and the integrated mechanism model was simulated by Simulink software to effectively query failure modes of the product. Secondly, the reversal forms of functions and standard solutions were added based on the substance-field model, and a method for obtaining failure events by reverse application of inventive principles in TRIZ (Teoriya Resheniya Izobreatatelskikh Zadatch) was proposed, forming a new path to obtain failure events. Finally, an AFD prediction process combining mechanism models and the reverse application of TRIZ problem-solving tools was established, and the failure prediction for a passive power-damping boring bar was conducted to verify the feasibility of this method. The proposed method improves the acquisition way of failure scenarios based on the traditional AFD process, effectively enhancing the rationality and accuracy of failure mode prediction, which has significant theoretical insights and practical engineering value for improving product design reliability.