The system of reflective double grating diffraction has been studied based on the optical theory. A mathematic model which is related to the reflective grating signal as well as the relative displacement has been established, and the characteristic of displacement is analyzed and simulated by Matlab, it is easy to find that the output light intensity of 0th signal come to the largest and the regulation of it stays stable. Besides, an ultra precision positioning system using double reflecting gratings is designed, the system chooses 0th grating signal as the control signal, through the closed-loop control by micro-computer, it can provide real-time detection of high-precision displacement and automatic location in ultra high precision. An intelligent proportion integration differentiation(PID) control technology based on the fuzzy radial basis function (RBF) neural networks has been designed in allusion to the characteristics such as non-linear and hysteresis of the reflect positioning system. With the affection both of them, the control parameter of PID system can be adjusted on time, the intelligent PID control system is able to realize the intelligentialize as well as to improve the adaptive ability. The result indicates that the intelligent PID control system has good interference immunity, high positioning efficiency, and the accuracy can achieve ±10 nm.