Abstract: In view of problem of big error between adjustment speed and expected speed of existing multistage belt speed-regualtion system adopting two-dimensional fuzzy control algorithm, a kind of adaptive neural fuzzy inference system (ANFIS) model was established, and a multistage belt speed-regualtion system based on ANFIS was designed. The speed-regualtion system takes instantaneous flow and real-time speed of the first belt as input, and takes regulating frequency of frequency converter as output to realize speed control. The Matlab simulation results show that the speed control error of the speed-regualtion system introducing ANFIS model is below 2%, capacity and belt speed matching rate are optimized, which has certain application value for energy saving and emission reduction of coal mine enterprises.