Specialized sensory hairs are important biological sensors for arthropods to detect and recognize environmental conditions including acoustic, pressure and airflow signals. However, the present design methodology of such biomimic micro devices are mainly depending on shape mimicking, which greatly restricts their performance. In this paper, a novel genetic algorithm based optimization model for design of piezoelectric functional hair is developed for improving its acoustic pressure or tactile sensitivity. Furthermore, the sensing mechanism of axially polarized piezoelectric hair is explored and the main influencing factors on sensitivity including hair configuration and axial strain distribution are determined. Then, a series of optimized hair configurations are obtained in a specific frequency band from 1 Hz to 500 Hz, whose average sensitivity of 2.21 × 10−3 V Pa−1 is 10 times greater than that of the straight hair of 2.15 × 10−4 V Pa−1 with the same size. For tactile load detection, the output voltage of the optimized hair is about 1.5 times as much as that of the straight hair. The obtained hairs are similar with the spider’s trichobothria and tactile hair, which presents an explanation of biological hairs sensitive to dynamic and static loads.