Аннотация:A mathematical model of building a neural network for diagnosing circulatory disordersin the vessels of a human’s circulation is considered. A procedure is proposed for creating a virtualdatabase for diagnosing with mathematical modeling of the blood flow in a quasi- one dimensionalapproximation of hemodynamics. Estimates of the accuracy of diagnosis are obtained using theproposed neural network for a set of circulatory pathologies. It is shown that some characteristicblood flow disorders reproduced in computation experiments are determined with fair reliability.