<p>Flexible sensing technologies for dynamic respiratory monitoring face critical limitations in environmental robustness and signal resolution accuracy. To address these challenges, a humidity-sensitive dielectric material was developed through intermolecular force modulation, synergistically integrated with a hermetically sealed digital mask to establish a medical-grade respiratory monitoring platform. A novel quantitative respiratory waveform analytical model was proposed, transcending conventional flexible sensors’ capability of merely tracking respiratory rhythms to enable precise quantification of pulmonary function parameters, including peak expiratory flow (PEF) and forced vital capacity (FVC). Leveraging a Darcy’s law-based porous media gas dynamics model, a linear response mechanism was identified between sensing signals and airflow/volume parameters (<i>R</i>² &gt; 0.995). Time–frequency characteristics of respiratory waveforms were extracted via synchrosqueezed wavelet transforms, revealing robust correlations between spectral signatures and physical activity intensity. Clinical validation in chronic obstructive pulmonary disease (COPD) cohorts demonstrated the system’s efficacy in detecting characteristic patterns of airway obstruction and diminished pulmonary elasticity, enabling early-stage diagnostics. Furthermore, a 1-dimensional convolutional neural network (1D-CNN) achieved high-accuracy cough event recognition (95.24% precision). A vertically integrated “sensing material–medical device–algorithm” framework is pioneered for home-based artificial intelligence (AI) respiratory disease management, advancing flexible electronics from physiological tracking to precision medical applications.</p> Graphical Abstract <p></p>