Abstract: Promoting high-quality employment is crucial for advancing new urbanization and achieving urban–rural integration and common prosperity. Based on data from the China Family Panel Studies (CFPS) from 2014 to 2022, this study constructs a multidimensional evaluation index system that integrates economic compensation, social security, job stability, subjective evaluation, and workload. By systematically comparing the performance of multiple linear regression, ensemble learning, and deep learning algorithms, the XGBoost model is employed to identify the core driving factors of employment quality among the rural transfer population, and SHAP values are applied to enhance interpretability. The results show that: 1) ensemble learning models exhibit clear advantages in predictive accuracy and generalization performance in identifying the driving factors of employment quality among the rural transfer population; 2) education level, industrial structure, income of other household members, labor union participation, and regional GDP are key determinants of employment quality, and their effects display stage-specific and threshold characteristics; 3) across different stages of economic development, the effects of investing in physical capital and investing in human capital evolve dynamically; 4) the core driving factors of employment quality follow a life-cycle evolution pattern with years of work experience and present spatial heterogeneity, with regions characterized by higher levels of marketization and more developed industrial systems exhibiting more mature human capital pricing mechanisms, while kinship, geographical, and industrial networks play a complementary role. Based on these findings, this study proposes policy recommendations, including the establishment of a targeted support system, the optimization of policy supply across different stages and levels, and the coordinated promotion of investing in physical capital and investing in human capital.