Human respiration is rich in physiological and pathological information for health assessment and illness prediction. In this work, an integrated triboelectric self-powered respiration sensor (TSRS) was developed for simultaneously monitoring human respiratory behaviors and NH3 concentration in exhaled gases. The TSRS based on Ce-doping ratio of 0.004 M exhibits a better selectivity and larger sensitivity of 20.13 ppm−1 compared with those synthesized in other doping ratios. Under the same moisture as the breathing gas (97.5%RH), the sensor holds greater sensitivity toward NH3 at low concentration (0.1–1 ppm) than that at high concentration (1–10 ppm), implying the capability in detecting trace level NH3 biomarker in human breathing gases. Furthermore, triggered by the expansion and contraction of chest during breathing, TSRS can spontaneously monitor human respiratory patterns and physiological process after physical exercise. Meanwhile, a theoretical modeling was established in terms of permittivity change under gas molecules adsorption and verified via finite element simulation to interpret the self-powered sensing behaviors. This research not only proposes a promising approach to develop an all-in-one wearable respiration monitoring system for personal health diagnosis but also sheds some light on the theoretical modeling of triboelectric self-powered gas-sensing in terms of permittivity change.