Abstract: In this paper, we studied on the minimum feedback bits number in multiple-input-single-output (MISO) systems with artificial noise assisted secure transmission. We used codebook to convey channel direction information quantized by random vector quantization (RVQ) method. We analyzed ergodic secrecy capacity loss quantitatively due to RVQ quantization considering different transmitted SNR and quantization error features. The closed-form of minimum feedback bits number was derived. To maintain a constant secrecy capacity loss, feedback bits should be scaled linearly with the logarithmic transmitted SNR as well as (Nt-1)(Nt is the transmit antenna number). Theoretical analysis and simulation results show that it satisfies security demands through adjusting feedback bits number according to our proposed closed form expression. Present research did not consider the case of low SNR. Our conclusion is suitable for all the SNR cases and makes contributions to the realization of artificial noise assisted physical secure transmission strategies.