Intelligent gait recognition system plays an important role in the field of identity recognition, physical training and medical diagnostics. In clinical medicine, no definitive diagnostic tool has been developed for the diagnosis of Parkinson's disease and hemiplegia. Thus, there is an urgent need to develop an effective and portable human -machine interaction system to monitor and recognize these symptoms. Herein, a self-powered strain sensor based on graphene oxide-polyacrylamide (GO-PAM) hydrogels is reported to monitor subtle human motions, including gait movements. The sensor can be used as a triboelectric nanogenerator (TENG) to collect mechanical energy. The output power of the TENG based on the 0.02 wt% GO-PAM hydrogel was up to 26 mW, which was 2.2 times that of the pure PAM hydrogel film. The capability of the TENG in powering electrical devices was demonstrated by lighting up 353 light-emitting diodes (LEDs) and powering an electronic thermometer. Besides, a wearable in -shoe monitoring system was designed which includes a flexible insole, a data processing module and a PC interface developed using Python. Among the models with different algorithms, the system with the artificial neural network (ANN) exhibits the highest recognition accuracy of 99.5 % and 98.2 % for human daily-life gait and pathological gait, respectively. This system provides a more convenient option for human gait monitoring and recognition, which can be used for a wide range of medical applications such as early diagnosis, rehabili-tation evaluation and treatment of patients.