Due to the harsh working environment of hydraulic excavators, the actuators are subjected to high intensity and long duration, resulting in fatigue-induced cracks, fractures, and other failures, leading to work interruptions and delays, reducing the operational efficiency of the excavator. Therefore, fatigue life prediction is of great significance in improving its safety, reliability, and operational efficiency. The load-time history at each articulation point of the working mechanism is the basis for fatigue life prediction. Due to the complex and variable forces experienced during the excavation process, it is difficult to obtain the working load spectrum through direct measurement, which can provide a more realistic prediction of the fatigue life. In this study, the excavation process of different materials under different working conditions of the excavator is studied to obtain a more accurate load spectrum and complete the life prediction of the working mechanism based on the discrete element principle. Taking the hydraulic excavator's actuators as the research object, the load spectrum of each articulation point under extreme conditions such as eccentric load and impact is obtained for different excavation materials, and the minimum life prediction of the boom is completed. Genetic algorithm (GA) and Self-Adaptive Fast Fireworks Algorithm (SF-FWA) are used for life prediction, and function fitting is performed using historical failure data for life calculation and comparative analysis. The analysis shows that SF-FWA is more effective than GA. The advanced algorithm provides more accurate predicted values for estimating fatigue life, but less accurate for estimating MTBF.