This study aimed to investigate the potential of magnetic particle imaging (MPI) to quantify artificial stenoses in vessel phantoms in near real-time. Custom-made stenosis phantoms with different degrees of stenosis (0%, 25%, 50%, 75%, and 100%; length 40mm, inner diameter 8mm, Polyoxymethylene) were filled with diluted Ferucarbotran (superparamagnetic iron-oxide nanoparticle (SPION) tracer agent, 500mmol (Fe)/l). A traveling wave MPI scanner (spatial resolution~2mm, gradient strength~1.5T/m, field of view: 65mm length and 29mm diameter, frequencies f1=1050Hz and f2=12150Hz) was used to acquire images of the phantoms (200ms total acquisition time per image, 10 averages). Standardized grey scaling was used for comparability. All measured stenoses (n=80) were graded manually using a dedicated software tool. MPI allowed for accurate visualization of stenoses at a frame rate of 5frames per second. Less severe stenoses were detected more precisely than higher-grade stenoses and came with smaller standard deviations. In particular, the 0%, 25%, 50%, 75%, and 100% stenosis phantom were measured as 3.7±2.7% (mean±standarddeviation), 18.6±1.8%, 52.8±3.7%, 77.8±14.8% and 100±0%. Geometrical distortions occurred around the center of the high-grade stenosis and led to higher standard deviations compared to lower grade stenoses. In the frame of this study the MPI signal depended linearly on the SPION concentration down to 0.05mmol (Fe)/l. Near real-time MPI accurately visualized and quantified different stenosis grades in vascular phantoms.