To solve the problems of accuracy and speed of the fifth set of RMB face value recognition, a design method of RMB face value recognition system based on MATLAB and image processing technology is given. The collected banknote images are pro- cessed by five modules: rotation correction, image preprocessing, image positioning, face value cutting and face value recognition. The experimental results show that the system can accurately calculate the total amount of money input while recognizing the face value of banknotes quickly in real time.