A tactile learning machine, which can simultaneously detect, memorize and thus learn the external stimuli, is of great interest with its potential in emerging wearable artificial intelligent neuromorphic electronics. Herein, we introduce an artificial tactile learning skin platform based on arrays of ferroelectric-gate field-effect transistors, enabling simultaneous sensing and learning of a variety of tactile information. Tactile pressure applied to a top-gate is received in terms of ferroelectric remnant polarization in a gate insulator, giving rise to analog conductance modulation dependent upon both magnitude and time of input pressure. Interestingly, our tactile learning skin with high-pressure reception sensitivity successfully mimics diverse synaptic functions, allowing a high-performance tactile perception learning. The 2-dimensional spatial learning and recognition are demonstrated, making them suitable for novel applications of wearable human-interactive electronics.