Abstract: Considering that three-degree-of-freedom artificial limbs based on the threshold control method are not intuitive and flexible and that pattern-recognized controllers with higher identification precision have a poor portability and usability, an online pattern-recognized control system based on programmable system-on-chip (PSoC) is proposed. The PSoC, a kind of low-power processor, is used as the main controller, and it is also used to design a portable four-channel surface electromyography (sEMG) signal acquisition system. A DP-PSO-SVM (dynamic programming-particle swarm optimization-support vector machine) algorithm is proposed for movement recognition, and then the prosthesis can be controlled online through a prosthetic actuator. Experimental results show that the classifier using the DP-PSO-SVM algorithm has a recognition rate of 96.7%, which is 4% higher than that of the classifier based on PSO-SVM algorithm; the controller exhibited an online recognition rate of 96.3% for six actions, and it meets real-time requirements.