Flexible tactile sensors have a sensitive material layer that is the core of perception. Based on functional materials, they can be categorized into piezoresistive, piezoelectric, and capacitive types. By employing laser engraving or photolithography techniques, biomimetic micro-pyramids and micro-column arrays are constructed to further enhance sensitivity. The surface micro-structured sensors developed by the Chinese Academy of Sciences can detect slight touches of 0.01N, approaching the human tactile threshold. In the field of surgical robotics, three-dimensional force feedback at the end of instruments is achieved through FBG optical fiber sensors. Specific packaging technologies are used to precisely encapsulate flexible substrates (medical-grade sensors utilize biocompatible materials). A dedicated electronic system is designed to convert and process the collected sensor signals in real-time, involving data fusion and AI algorithm analysis. By integrating multidimensional data such as pressure, temperature, and vibration, multimodal signal fusion is performed, while simultaneously analyzing tactile and optical data to identify tissue types. Convolutional Neural Networks (CNN) are employed to analyze tactile texture features, or Transformer models are used to predict object slip trends. For instance, surgical robots optimize grasping strategies through online learning. The technological breakthroughs in flexible tactile sensors deeply integrate “biomimetic perception” and “embodied intelligence,” advancing surgical robots from being mere “execution tools” to an integrated system of “perception-decision-execution.”