Abstract: To solve the problems of feedback lag, nonlinear correlation and interference factors in the dosing process and the problems of the traditional proportion integration differentiation（PID）controller with slow response speed, high secondary disturbance and weak anti-interference ability, a new intelligent sodium hypochlorite dosing scheme was designed to meet the requirements of chlorine concentration and dosage of effluent in the chlorination process for water purification. In the new intelligent chlorination scheme, ControlLogix system was used as the lower machine, and model predictive control（MPC） module was embedded in the lower machine. In the model predictive control module, the historical data of the water plant was used to build the prediction model, and the model training was carried out in the simulation practical application. In the control module, the empirical formula of the water plant was integrated to realize the intelligent control of the dosing system. The results show that the residual chlorine concentration is stable at the range of 0.97-1.16 mg·L^-1. Compared with the traditional chlorination scheme, the consumption of sodium hypochlorite solution per hour in the new intelligent chlorination scheme is reduced by 0.5%-1.0%.