Piezoelectric actuators are widely employed in the field of micro-/nanomanipulation. However, hysteresis is the dominant issue in piezoelectric actuators, which leads to a great challenge to achieve high precision micromanipulation. Proportional-integral-derivative (PID) control is an efficient approach to reduce hysteresis effect in piezoelectric actuators. However, its parameter tuning is a time-consuming work for PID motion tracking control implementation. In this work, the neural networks (NN) is adopted to provide a functional model for PID with optimized parameters. It enables an intelligent and adaptive motion tracking process. The effectiveness of the presented NN-based PID control scheme is verified by performing simulation studies.