Abstract: At present,many of the looms uses conventional PID control in China,which relays on the mathematical model and can′t get a good control result.Because of conventional PID control′s defects,an adaptive PID control algorithm based on RBF neural network with a Kalman filter is designed.The control algorithm uses three input-single output RBF radial neural network which learns system′s performance to find the best combination of PID,uses a Kalman filter which effectively filters out various noises of the loom,and achieves a constant tension value.The simulate results indicate that control effect and dynamic performance for adaptive PID control for warp tension system based on neural network are obviously superior to those of conventional PID control.