In the field of wearable health monitoring, asymmetric adhesion presents a core challenge in achieving long-term comfort and safety. Conventional symmetric adhesion materials are prone to contaminant accumulation due to their strong adhesion across the entire surface. This study constructs an asymmetric adhesion interface featuring covalent strong adhesion on the top surface (53.2 kPa) and anti-adhesion on the spontaneously settling bottom surface (2 kPa) within a poly(acrylic acid)-chitosan system, utilizing an innovative dynamic chemical gradient self-assembly strategy. Furthermore, we developed a conductive hydrogel designed as a hydrogel strain sensor, exhibiting exceptional mechanical properties (2135 %) and high electrical conductivity (1.2 S/m), enabling the monitoring of various motor activities in the human body. This hydrogel can also function as a skin surface electrode for long-term, high-fidelity ECG and EMG signal capture. Additionally, we prepared a hydrogel-friction electric nanogenerator that serves dual purposes: energy harvesting and pressure sensing. An integrated flexible smart insole system was also designed, demonstrating outstanding performance in monitoring and diagnosing Parkinson's symptoms. Among various models employing different algorithms, the convolutional neural network achieved the highest accuracy (97.4 %) in recognizing Parkinson's gait. This system offers a novel tool for monitoring and recognizing Parkinson's gait, facilitating the tracking of rehabilitation processes and personalized treatment planning, thus showcasing its potential for groundbreaking applications in wearable medical monitoring.