Abstract Bio-inspired near-sensor computing, which integrates sensing and processing functions, presents a promising strategy to enhance efficiency and reduce latency in such applications. Here, we introduce tactile sensory nerve systems with biologically realistic energy efficiency, utilizing starfish-inspired capacitive pressure sensors integrated with flexible memristors. These starfish-inspired sensors, with their high aspect ratio (~3) and stress-focusing, hourglass-shaped dielectric microstructures, enable highly sensitive tactile detection across a broad pressure range, effectively mimicking the properties of human skin. Artificial tactile sensory nerves, which integrate the capacitive sensor with a flexible memristor exhibiting synaptic plasticity, function reliably as energy-efficient near-sensor computing systems by bio-realistically transducing mechanical stimuli into transient electrical signals. The developed system operates as both an artificial nociceptor and a tactile near-sensor computing unit, with energy consumption approaching biological levels at approximately 140 pJ and 2.2 fJ, respectively. This neuro-inspired localized computing strategy offers a physical platform for advanced smart user interface applications.