Fatigue failure under alternating loads represents a significant failure mode for automotive components. This study focuses on predicting the fatigue life of automotive parts subjected to multi-level stress loading. Current nonlinear models often require extensive test data or face challenges in selecting appropriate reference values, which imposes limitations on automotive reliability assessments. By analyzing the process of fatigue damage transformation between two levels of stress, the relationship, i.e., $S_i^2\times \mathrm{\Delta }{S}_i\text{=}{S}_{i+1}^2\times \mathrm{\Delta }{S}_{i+1}$ , considering the equivalent transformation of adjacent stresses based on the S-N characteristic curve of materials. The equivalent formula for cumulative fatigue damage and the expression for residual fatigue life between adjacent stresses acress two, three, and higher stress levels are derived. Then a fatigue life prediction model based on the stress equivalence relation is proposed. The calculation process of this model relies solely on the fatigue life of the materials subjected to at least two stress levels to determine the S-N characteristic curve. The experimental data utilized include existing test results for two-, three-, four-, and five-level stress loading, as reported by relevant scholars. Specifically, the two-level stress materials analyzed are AL-2024 and 30CrMnSiA, the three-level stress material is LY12CZ, and the four- and five-level stress material is an aluminum alloy. The average and maximum relative errors of the Miner model, Manson model, Subramanyan model, Hashin model, and the proposed model are all calculated and compared. Additionally, the statistical values of cumulative fatigue damage, cumulative fatigue life, and the discrepancies between cumulative fatigue damage and test cumulative fatigue damage predicted by each model are summarized. The results show that the proposed model, which is based on the equivalent transformation of adjacent loads, demonstrates superior overall performance in predicting fatigue life and damage compared to the Miner model, Manson model, Subramanyan model, and Hashin model, thereby offering a more accurate application for fatigue life and damage prediction under multi-level loading conditions.