Electronic stethoscope used to detect cardiac sounds that contains essential clinical information is a primary tool for diagnosis of various cardiac disorders. However, the linear electro‐mechanical constitutive relation makes conventional piezoelectric sensors rather ineffective to detect low‐intensity, low‐frequency heart acoustic signal without the assistance of complex filtering and amplification circuits. Herein, we find that triboelectric sensor features superior advantages over piezoelectric one for micro‐quantity sensing originated from the fast saturated constitutive characteristic. As a result, the triboelectric sensor shows ultrahigh sensitivity (1215 mV/Pa) than the piezoelectric counterpart (21 mV/Pa) in the sound pressure range of 50 – 80 dB under the same testing condition. By designing a trumpet‐shaped auscultatory cavity with a power function cross‐section to achieve acoustic energy converging and impedance matching, triboelectric stethoscope delivers 36 dB signal‐to‐noise ratio for human test (2.3 times of that for piezoelectric one). Further combining with machine learning, five cardiac states can be diagnosed at 97% accuracy. In general, the triboelectric sensor is distinctly unique in basic mechanism, provides a novel design concept for sensing micromechanical quantities, and presents significant potential for application in cardiac sounds sensing and disease diagnosis.This article is protected by copyright. All rights reserved