Abstract: The dosing amount of agent in Buliangou Plant′s coal slurry water treatment process cannot be adjusted in a real time manner due to heavy reliance on experience of operators and unavailability of monitoring system, affecting, therefore, observation and evaluation of slurry settling and thickener′s demixing affects. To address the issues, an intelligent agent dosing system is developed. The system is comprised of the data collection layer, data layer, logic control layer and intelligent decision-making layer. Coupled with the use of flowmeter, densometer, slime-water interface meter and turbidometer, the system can execute collection of thickener′s key data, such as its feed rate, feed concentration, overflow turbidity and height of clarified layer as well as real-time monitoring and control of running state of agent dosing devices with the help of the system′s display interface. And intelligent prediction of agent dosing amount can be made with the intelligent agent dosing model constructed based on TCN-BP algorithm. Field application of the system shows that the values measured by the system′s meters are in good agreement with the manually measured values, and the absolute values measured by the densometer and slime-water interface meter differ with the manually measured values by 6 g/L and 0.12 m, respectively, on average; and the use of the system has led to possibility to make intelligent adding of agents, effective reduction of consumption of agents, a save of cost of agents amounting to 165 400 yuan a year, improved slurry water treatment efficiency and cut of labor cost for only one operator is required instead of two, thus laying a solid foundation for intelligent development of the plant.