For the optimal control problem of predinitrification wastewater treatment process, a Hopfield neural network optimization method based on the Lagrange multiplier is proposed. Firstly, under the constrain of some key effluent pollutant qualities, a wastewater treatment optimization objective function is constructed to minimise the energy consumption. Then, the set points in bioreactor of both dissovled oxygen concentration in the 5th compartment and nitrate concentration in the 2nd compartment are optimized by Hopfiled neural network, respectively. Both concentrations are controlled by PID controller. Finally, based on the international standard benchmark, the simulation results show that through the optimization of Lagrange multiplier Hopfield neural network, the energy consumption of wastewater treatment process is reduced obviously under constraints of effluent pollutant qualities.