Given the technical constraints of improving the sensitivity, stability and mechanical flexibility of current flexible resistive pressure sensors for complex curved-surface applications, this paper proposes a design strategy based on a neuron-like conductive network structure. First, a hybrid conductive network of carbon nanotubes/silver nanowires is built to model the distributed connection structure of biological neurons. Then, by using percolation theory models and finite-element simulations, a multi-scale piezoresistive response model will be established. Next, a multi-objective genetic algorithm will be employed to optimize the best ratio of carbon nanotubes/silver nanowires (CNTs/AgNWs) at a mass ratio of 1:2 and a concentration of 1.5 wt%. Finally, spin coating and then curing will be used for the sensor. Based on the experimental results, the sensor has a sensitivity of 30.4 kPa⁻¹ in the low-pressure area (0-2 kPa); after 10,000 cycles, there is only a 2.9% reduction in performance, and it maintains about 86.3% of its original performance at a 5 mm curvature. This study has verified that the biomimetic network structure effectively addresses the problem of synergistic optimisation for multiple performance indicators of flexible sensors through dynamic reconstruction of conductive pathways and stress dispersion mechanisms.
