Abstract: Finite control set model predictive control (FCS-MPC) has great advantages in dealing with optimization problems, but its control accuracy declines sharply at low sampling frequency. Taking three-phase voltage source rectifier (VSR) as an example, the essence of traditional solutions for this problem, mostly the extended vector or modulated MPC (M-MPC), is to improve the control accuracy by increasing the switching frequency, which can not achieve an optimization effect. To solve the problem of poor performance of FCS-MPC under low sampling frequency and improve the control accuracy with the minimum increase of switching frequency, a new control set expansion method for three-phase VSR is proposed based on the analysis of the nonlinear relationship between switching frequency and control error of FCS-MPC. The control set is expanded in the proposed method by constructing 12 new voltage vectors and by using these new added voltage vectors. The control accuracy of FCS-MPC at low sampling frequency is significantly improved and the increase of switching frequency is minimized. The proposed method could also achieve better performance than M-MPC under the same switching frequency. In addition, the expansion mode could be changed easily by adjusting the weighting factor of the added vectors, which retained the characteristics of simple and direct of FCS-MPC. The effectiveness of the proposed method is verified by simulation and experiment.