The underflow concentration of the thickener in the coal preparation plant affects the operational safety of the thickener and the efficiency of the filtration system. Predicting the underflow concentration and using it for the intelligent control of the coal slime water system is of great significance. Based on the basic laws of solid settling and mass balance, a prediction algorithm for the underflow concentration and the settling velocity of coal slurry particles with different particle sizes in the thickener was constructed using online instantaneous and historical data including the flow rate and concentration of the thickener feed, the dosage of coagulant, and the flow rate and concentration of the thickener underflow, as well as offline data including coal slurry particle size distribution and optimal dosage. The algorithm was applied in a coal preparation plant to verify its accuracy. The results showed that without adding flocculant, particles <0.045 mm in size in the thickener would circulate in the system with the overflow water, while coal slurry with a particle size range of 0.045-0.125 mm would remain in the thickener as suspended solids, and particles larger than 0.125 mm would settle to the bottom; After adding medication, coal slurry of various particle sizes settles in the form of flocs. The underflow concentration prediction algorithm based on settling rate and mass balance can accurately predict the bottom flow concentration, providing a basis for decision-making in coal slurry intelligent systems.