Abstract: To address key issues in load spectrum compilation for tractors, domestic and international research progress was reviewed, and data collection, pre-processing, and load extrapolation techniques were analyzed, with a focus on time-domain extrapolation and rainflow extrapolation methods．Research results showed that load spectrum compilation for agricultural tractors faced challenges such as large data collection volumes and a lack of high-horsepower data．Introduction of non-parametric estimation and deep learning models has significantly improved load spectrum compilation accuracy and efficiency．It is recommended to combine big data and deep learning technologies to optimize compilation process, enhancing design reliability and service life of tractors and their components．