Abstract: PID neural network (PIDNN) control algorithm is suitable for the decoupling control of nonlinear multivariable coupled system, because of its better dynamic and steady performance, and stronger decoupling and disturbanceattenuate ability. By updating weights via Quantum Particle Swarm Optimization(QPSO), this paper further proposes a multivariable PID neural network control algorithm. The simulation results show that the present algorithm can attain better decoupling control effect and stronger disturbanceattenuate ability.