The sound generated by a reed-type artificial vocal fold was predicted by a one-mass modeling and numerical flow simulation to examine the sound generation mechanisms of the artificial vocal fold. For the one-mass modeling, the reed oscillation was modeled with an equivalent spring constant, and the flow rate was estimated by Bernoulli’s equation. For the flow simulation, the flow and acoustic fields were predicted with compressible Navier-Stokes Equations, while the reed oscillation was calculated by a one-dimensional beam equation. The experimentation was conducted by measuring the sound of an artificial vocal fold in an anechoic chamber. The results of the acoustic measurement showed that the sound amplitudes in the flow simulation agreed well with the experiment, while the one-mass model underestimated the amplitudes in a higher frequency range. Reed displacement and flow rate comparisons indicated that the flow retention in the reed retainer caused the asymmetry in the flow rate waveform, hence producing larger amplitudes for the flow simulation in the higher frequency range. The flow simulation enabled to predict this flow retention which cannot be modeled in the one-dimensional one-mass model, and it is anticipated to apply the flow simulation to develop a better artificial vocal fold.