Hemodialysis is the most common treatment for patients with end-stage renal disease. For hemodialysis, consistently functional vascular access must be surgically created with an anastomosis of artery and vein, referred to as an arteriovenous fistula (AVF). However, AVF dysfunction may occur over time. Angiography and Doppler ultrasound are usually used to detect the flow or the diameter of the AVF. But they require well-trained operators and are expensive, and even angiography is invasive. In this study, a noninvasive approach based on stethoscope auscultation for monitoring AVF stenosis was proposed. Here, a wireless blood flow sound recorder was designed to record blood flow sounds wirelessly. In order to effectively extract the varying feature of blood flow sounds for AVF stenosis, the 2-D feature pattern built from S-transform was also proposed as the feature in the AVF stenosis detecting algorithm. Different from other frequency-related coefficients, the feature pattern can contain the information of blood flow sounds in time and frequency domains simultaneously. Preliminary findings showed that the proposed approach can provide high-quality estimation of AVF stenosis (positive predictive value = 87.84% and sensitivity = 89.24%).