The cognition of spatiotemporal tactile stimuli including fine spatial stimuli and static/dynamic temporal stimuli is paramount for intelligent robots to feel their surroundings and complete manipulation tasks. However, current tactile sensors have restrictions in simultaneously demonstrating high sensitivity and performing selective responses to static/dynamic stimuli, making it a challenge to effectively cognize spatiotemporal tactile stimuli. Here, we report a high-sensitive and self-selective humanoid mechanoreceptor (HMR) that can precisely respond to spatiotemporal tactile stimuli. The HMR with PDMS/chitosan@CNTs graded microstructures and polyurethane hierarchical porous spacer exhibits high sensitivity of 3790.8 kPa-1. The HMR demonstrates self-selective responses to static and dynamic stimuli with mono signal through the hybrid of piezoresistive and triboelectric mechanisms. Consequently, it can respond to spatiotemporal tactile stimuli and generate distinguishable and multi-type characteristic signals. With the assistance of the convolutional neural network, multiple target objects can be easily identified with a high accuracy of 99.1%. This work shows great potential in object precise identification and dexterous manipulation, which is the basis of intelligent robots and natural human-machine interactions. Video Supporting Video 1.mp4 Supporting Video 2.mp4 Supporting Video 3.mp4 Supporting Video 4.mp4