The induced radioactivity in stereotactic body radiation therapy with a flattening-filter-free 10 MV beam model (10 FFF SBRT) was investigated for the risk to therapists. Methods: This study was performed on a Varian TrueBeam linac. The induced radioisotopes were identified by γ spectroscopy. The dose rate from the induced activity was measured for 12 treatment cycles in 4 h continuously. The impacts of the characteristic factors of 10 FFF SBRT on the dose rate were investigated, including monitor units (MU), beam rate, field size, and flattening filter. The dose rate was compared between the SBRT plans and conventional fractionation plans. A mathematical model was used to analyze the results and estimate the annual dose to therapists. Results: (a) The induced radioisotopes included 24 Na, 28 Al, 38 Cl, 56 Mn, 66 Cu, 187 W, and 196 Au. (b) In 4 h, the total dose contribution ratios were more than 70% for 28 Al, about 20% for 56 Mn, and 10% for all other long-lived radioisotopes, combining doses at the isocenter and end of the treatment couch. (c) The dose rate showed a nonlinear growth with increasing MU and beam rate. The variation of the dose rate was complicated with the jaw field and not sensitive to the MLC field. The removal of the flattening filter reduced the dose rate by about 40%. The dose level of SBRT was two to three times that of conventional fractionation. (d) The estimated annual dose to therapists was up to 0.20 mSv/y. Conclusions: The induced radioactivity in 10 FFF SBRT was higher compared with that in 10 MV conventional fractionation. More MU and higher beam rate were the primary factors that caused the increase. The therapists should wait longer after beam-off to reduce the occupational dose. In addition, aluminum and manganese should be less used in the treatment room.