This paper uses the training of OC-SVM to analyze the mechanism of the Quantum-behaved particle swarm and develops a method of training OC-SVM based on the directional- QDPSO .The new position of the directional particle is calculated based on the current global best point(gBest), which identified the optimized direction conforms to Zoutendijk fastest decline method principle.In the initialization, the position of one particle is initialized according to SMO, which makes its position nearer to the global optimum solution. The boundary points of subjected plane are concerned as the initialized position of other particles, so as to make the searching area wider.The experiment result shows that the convergence and the generalization of D-QDPSO is good, the misrecognition of D-QDPSO is 0.12% lower than that of SMO, and the operating speed is 2 times faster than that of LPSO.