This study aims to deeply investigate the mechanism of NHC-Mn catalyzed ester hydrosilylation reaction and the impact of different ligands on reaction activity through the comprehensive application of DFT calculations and machine learning technologies. We initially proposed four possible reaction mechanisms including two radical mechanisms and two 2e mechanisms. Then calculate the potential energy surfaces of them. It was demonstrated through calculations that the radical mechanisms were implausible, while the 2e outer-sphere mechanism was proven to be reasonable. A detailed analysis of the Si–H bond activation step in different reaction mechanisms highlighted the crucial role of steric effects in the process. Furthermore, the rate-determining step in the 2e mechanism (outer-sphere), the hydride transfer process, was also thoroughly studied. Eventually, a predictive model was successfully constructed using machine learning methods, which is a model based on the Ridge algorithm with 17 features that accurately predict the impact of different types of ligands on the reaction barrier. Analysis of the model revealed important factors affecting reaction activity, such as the vertical ionization potential of the metal hydride intermediate, the charge on the metal hydride, and the bond order of the Mn–H bond.