The development of bioinspired gradient hydrogels with self-sensing actuated capabilities for remote interaction with soft-hard robots remains a challenging endeavor. Here, we propose a novel multifunctional self-sensing actuated gradient hydrogel that combines ultrafast actuation and high sensitivity for remote interaction with robotic hand. The gradient network structure, achieved through a wettability difference method involving the rapid precipitation of MoO2 nanosheets, introduces hydrophilic disparities between two sides within hydrogel. This distinctive approach bestows the hydrogel with ultrafast thermo-responsive actuation (21° s-1) and enhanced photothermal efficiency (increase by 3.7 °C s-1 under 808 nm near-infrared). Moreover, the local cross-linking of sodium alginate with Ca2+ endows the hydrogel with programmable deformability and information display capabilities. Additionally, the hydrogel exhibits high sensitivity (gauge factor 3.94 within a wide strain range of 600%), fast response times (140 ms) and good cycling stability. Leveraging these exceptional properties, we incorporate the hydrogel into various soft actuators, including soft gripper, artificial iris, and bioinspired jellyfish, as well as wearable electronics capable of precise human motion and physiological signal detection. Furthermore, through the synergistic combination of remarkable actuation and sensitivity, we realize a self-sensing touch bioinspired tongue. Notably, by employing quantitative analysis of actuation-sensing, we realize remote interaction between soft-hard robot via the Internet of Things. The multifunctional self-sensing actuated gradient hydrogel presented in this study provides a new insight for advanced somatosensory materials, self-feedback intelligent soft robots and human–machine interactions.