The most common treatment for end-stage renal disease (ESRD) patients is the hemodialysis (HD). For this kind of treatment, the functional vascular access that called arteriovenous fistula (AVF) is done by surgery to connect the vein and artery. Stenosis is considered the major cause of dysfunction of AVF. In this study, a noninvasive approach based on asynchronous analysis of bilateral photoplethysmography (PPG) with error correcting output coding support vector machine one versus rest (ESVM-OVR) for the degree of stenosis (DOS) evaluation is proposed. An artificial neural network (ANN) classifier is also applied to compare the performance with the proposed system. The testing data has been collected from 22 patients at the right and left thumb of the hand. The experimental results indicated that the proposed system could provide positive predictive value (PPV) reaching 91.67% and had higher noise tolerance. The system has the potential for providing diagnostic assistance in a wearable device for evaluation of AVF stenosis.