A novel, non-invasive method to diagnose bladder outlet obstruction involves the recording of noise with a contact microphone pressed against the perineum (between anus and scrotum). This noise results from flow-generated vortices caused by prostatic obstruction. We developed a computational fluid dynamic (CFD) urethral model including urethral geometry to study the relation between generated noise and the degree of obstruction. This model comprised a bladder, bladder neck, prostate and urethra. Calculations were carried out at four bladder pressures, five degrees of obstruction and three obstruction shapes. For each of the sixty simulations, the velocity and pressure distributions along the urethra were calculated including wall shear stresses to localize flow transition from disturbed to normal. Negative pressures at the obstruction outlet induced recirculation of flow. The location of transition was independent of the applied bladder pressure, but it depended primarily on the degree and secondarily on the shape of the obstruction. Based on the presented results, we hypothesize that the location of the maximum amplitude of perineal noise mainly depends on the degree and shape of the prostatic obstruction. Our future aim is to test our hypothesis in male patients and to extend the presented model to 3D with a viscoelastic urethral wall to calculate the fluid–wall interaction.