Transcranial magnetic stimulation (TMS) stands as a widely employed neuromodulation technique for addressing various brain disorders. The hand motor hotspot (hMHS) holds particular significance in TMS applications, serving to ascertain personalized treatment targets and stimulation intensity. However, the conventional determination of hMHS remains time-consuming. Our objective was to expedite the identification of hMHS solely based on neuroimaging data. In this investigation, we pinpointed the hMHS on the cortical surface in depressed patients utilizing TMS-derived data and magnetic resonance imaging (MRI) data. Employing the kernel density estimation method, we developed a probability map for hMHS, subsequently utilizing a machine learning (ML) model to discern subjects suitable for the probability map application. The hMHS probability map was established, and the vertex with the highest probability was designated as the group hMHS. For subjects closely aligning with the group hMHS, direct application of the hMHS probability map was feasible. Conversely, for other subjects, our ML model, trained on cortical structure data of the sulci, could identify them. Our method achieved an 88% accuracy in hMHS determination and, when compared to traditional methods, exhibited an average time-saving of approximately 50%. In conclusion, our proposed method offers an efficient and rapid solution for hMHS identification during TMS treatment.