To understand the neural mechanism of repetitive transcranial magnetic stimulation (rTMS), the after-effects following one session or multiple days of stimulation have been widely investigated. However, the relation between the short-term effect (STE) and long-term effect (LTE) of rTMS is largely unknown. This study aims to explore whether the after-effects of 5-day rTMS on supplementary motor area (SMA) network could be predicted by one-session response. A primary cohort of 38 healthy participants underwent five daily sessions of real or sham continuous theta-burst stimulation (cTBS) on the left SMA. Resting-state functional magnetic resonance imaging (fMRI) data were acquired at the first (before and after the first stimulation) and sixth experimental day. The SMA connectivity changes after the first cTBS and after 5 days of stimulation were defined as STE and LTE, respectively. Compared to the baseline, significant STE and LTE were found in the bilateral paracentral gyrus (ParaCG) after real stimulation, suggesting shared neural correlates of short-and long-term stimulations. Region-ofinterest analysis indicated that the resting-state functional connectivity between SMA and ParaCG increased after real stimulation, while no significant change was found after sham stimulation. Leave-one-out cross-validation indicated that the LTE in ParaCG could be predicted by the STE after real but not sham stimulations. In an independent cohort, the after-effects of rTMS on ParaCG and short-to long-term prediction were reproduced at the region-of-interest level. These imaging evidences indicate that onesession rTMS can aid to predict the regions responsive to long-term stimulation and the individualized response degree.